Streamline PostgreSQL Projects with Vollna

We are looking for an experienced full-stack developer to develop and integrate a REST API using Spring Boot and connect it with our existing Angular application.

Project Requirements:
Develop secure and scalable REST APIs using Spring Boot
Integrate APIs with Angular frontend application
Implement CRUD operations (Create, Read, Update, Delete)
Ensure proper communication between frontend and backend
Connect APIs with database (MySQL/PostgreSQL)
Debug and fix any integration issues if needed
Required Skills:
Strong experience in Java & Spring Boot
Experience with Angular frontend development
REST API design and integration
Database handling (MySQL / PostgreSQL)
Understanding of authentication (JWT is a plus)
Deliverables:
Fully working backend APIs
Integrated Angular frontend with APIs
Clean and well-structured code
Basic documentation for setup and usage

We are looking for someone who can deliver clean, maintainable code and complete the integration smoothly.

Project Overview
I am seeking a developer to build a custom AI platform to replace a third party tool currently in use. The platform will power a sports specific AI assistant designed for coaches athletic directors and sports organizations. The goal is to create a reliable scalable system with full ownership and control over functionality data and monetization.

The platform must support centralized training multiple deployments with custom configurations subscription billing usage limits and a serverless architecture. It should also be designed to support future mobile app integration.

Core Objectives

* Build a production ready AI system with strong performance and reliability
* Enable monetization through subscription and usage based pricing
* Support multiple customer specific deployments with customized training
* Eliminate reliance on external AI agent platforms

Technical Stack Preferred

* Backend Node.js with serverless functions using Vercel or AWS Lambda
* AI integration with OpenAI GPT models and embeddings
* Database PostgreSQL with preference for Supabase
* Vector database Pinecone Weaviate or Supabase vector
* Payments integration using Stripe
* Frontend Phase two React or Next.js

Developers may recommend alternatives but should explain tradeoffs.

Key Features and Requirements

AI Agent Core

* Integration with OpenAI chat models
* System prompt for sports specific intelligence
* Retrieval augmented generation using embeddings
* Support for future tool or function calling

Multi Deployment Architecture

* Each deployment should have a unique identifier
* Support custom prompt layer per deployment
* Allow additional training data specific to that deployment
* Be embeddable via a lightweight JavaScript snippet

Training System

* Ability to upload and manage training content
* Store embeddings for both global and deployment specific knowledge
* Retrieve relevant documents dynamically during queries

Usage Tracking and Limits

* Track usage per user either by interaction or tokens
* Enforce plan limits
* Store detailed usage logs

Stripe Integration

* Support subscription plans monthly or annual
* Multiple pricing tiers with defined usage limits
* Handle webhook events including subscription creation updates and cancellations

Embed System

* Provide a JavaScript snippet to embed the AI assistant on external websites
* Support deployment specific configuration
* Ensure secure handling of user sessions

Serverless Architecture

* Fully serverless backend
* Optimized for scalability and performance
* Clean API design

Mobile Ready API

* Backend must support future mobile applications
* Stateless architecture preferred

Database Requirements High Level

* Agents for global system configuration
* Deployments for customer specific setups
* Training documents with embeddings
* Users and subscriptions
* Usage logs

Admin and Management Capabilities

* Create and manage access plans and pricing
* Invite and manage users
* Assign and update user access plans
* View user conversation history
* View uploaded training files per user
* Create and manage deployments

Deliverables

* Fully functional backend system ready for deployment
* Database schema and migrations
* API endpoints for chat training ingestion usage tracking and deployment management
* Complete Stripe integration
* Basic embed script
* Documentation for setup deployment and usage

Nice to Have

* Admin dashboard interface
* Streaming responses
* Role based access control
* Analytics for usage and engagement

Performance Expectations

* Fast response times typically under two to three seconds
* Reliable uptime
* Clean maintainable code
* Scalable architecture

Developer Requirements

* Experience working with OpenAI or similar AI systems
* Experience building retrieval augmented generation systems
* Strong backend development skills
* Experience with Stripe subscriptions and billing
* Familiarity with vector databases

Project Scope and Timeline

* Phase one MVP expected within one to three months
* Phase two enhancements and interface development to follow

Proposal Instructions

* Include relevant experience and examples of similar work
* Provide links to past projects or repositories
* Share recommended technical approach if different
* Include estimated timeline and cost
* Suggest any improvements to the architecture

Goal
Build a scalable sports intelligence AI platform that can be monetized and deployed across multiple customer segments

• Describe a production AI system you have built using the OpenAI API or similar. Include what you personally built and the scale of usage.
• Explain how you have implemented retrieval augmented generation including how you stored embeddings and retrieved relevant context at runtime.
• How would you design a system that supports both global training and deployment specific training while keeping data isolated between customers?
• How would you implement subscription billing and enforce usage limits using Stripe for an AI product?
• If you were starting this project today what would you build first and how would you sequence the work over the first 4 to 8 weeks?

We are GK Bikes, a fast-growing company in the bicycle and electric mobility space in India. We are looking to build a **world-class AI voice calling system** for our dealer network focused on **payment collections, follow-ups, and relationship management**.

This is not a basic bot project. We are building a **production-grade AI collections agent** that can handle real dealer conversations in **Hindi / Hinglish / English** and integrate deeply with our CRM.

We are looking for someone who has already worked on **AI voice agents in real-world environments**, not someone learning on the job.

---

**Scope of Work:**

You will be responsible for designing and implementing an AI voice calling system that can:

1. Make outbound calls to dealers for payment follow-ups
2. Hold natural conversations (Hindi/Hinglish/English)
3. Capture responses and update CRM
4. Extract structured data:

* Payment commitment (amount, date, mode)
* Reason for delay
* Sentiment / confidence level
5. Schedule follow-ups automatically
6. Escalate calls (transfer to human when needed)
7. Maintain conversation history per dealer

---

**Tech Stack (Preferred):**

* Voice AI: Vapi AI or Retell AI (or equivalent high-end platform)
* Telephony: Exotel (India-based calling compliance)
* Backend: Node.js / Python
* CRM Integration: API-based (custom CRM)
* Database: PostgreSQL / Firebase / similar

---

**Key Requirements:**

* Experience building AI voice agents (not just chatbots)
* Strong understanding of real-time voice pipelines (STT + LLM + TTS)
* Experience integrating telephony systems (Exotel, Twilio, etc.)
* Ability to design conversational logic (not just code)
* Experience with multilingual systems (Hindi preferred)
* Ability to handle latency, interruptions, and real-world call behaviour

---

**Bonus (Highly Preferred):**

* Experience with Indian market / users
* Experience in collections / fintech / recovery systems
* Understanding of Hinglish conversational patterns
* Experience with call analytics and outcome tracking

---

**What We Already Have:**

* Defined conversation flows
* Knowledge base and scripts
* CRM structure and dealer data
* Clear business logic and use cases

---

**What We Expect From You:**

* Suggest best architecture (we are open to improvements)
* Build a robust, scalable system
* Deliver production-ready solution (not prototype)
* Ensure high call quality and low latency
* Work closely with us on iterations

---

**Project Type:**
Initial build with long-term engagement possible

---

**How to Apply:**

Please include:

1. Links to similar AI voice projects you’ve built
2. Tech stack you used
3. Your approach to handling Hindi/Hinglish conversations
4. How you would structure this system (briefly)

---

**Important:**
We are looking for **top 1% talent**. If you have not built real AI voice calling systems before, this is not the right fit.

---

We are building something category-defining in the Indian dealer ecosystem. If you’re serious about AI systems at scale, this will be a high-impact project.

• Describe your recent experience with similar projects
• Describe common challenges of maintaining SLAs
• How do you operationalize your projects?

READ THE FULL JOB POST BEFORE APPLYING.

We’re hiring a Full Stack Engineer to own Open Letter Connect (OLC) — a live B2B SaaS platform used to run high-volume direct mail, data, and automation workflows. This is a hands-on product owner + architect + builder role.

What You’ll Own:
- UI/UX and backend
- Product architecture, performance, and uptime
- Feature delivery and technical roadmap
- Code quality and releases
- Working with a small team
**When it breaks or ships — it’s yours.**

Tech Stack you MUST be strong in:
- Node.js
- React
- Laravel
- PostgreSQL
- Server and performance optimization

AI Experience Required:
- Designing and deploying AI agents and bots
- Using LLMs for automation and workflows
- Acting as an AI architect, not just calling APIs

Who We Want
You’ve:
- Built or led a real B2B SaaS product
- Owned production systems with paying customers
- Led a small dev team
- Cared deeply about UX, not just code
- You are a passionate coder and have the ability to dive in
- CRM, automation, or marketing-tech experience is a big plus.

To Apply:
1. Answer ALL the screening questions (We only review applicants who include a link, demo, or screenshot of a real product they’ve built or owned that is currently (or was previously) used by paying customers.)
2. Add a cover letter about why you are interested in this job and include the word "HAPPY" so I know you read this post fully.
3. DO NOT APPLY IF YOU: A) work with an agency, B) don't have the necessary experience, including our tech stack, C) can't work 9-5 EST Mon-Fri, D) don't hold yourself to a high standard

• Send a link or screenshot of a real SaaS product you built or owned. Tell us exactly which parts you were responsible for.
• Show us one workflow or feature you designed that you’re proud of. Why are you proud of it?
• Explain one example of how you currently use AI to help with your development workflow (writing prd's, coding, qa, etc.) and provide one example of when you won't use AI.
• Show us the ugliest part of a system you built and why you feel this way (add screenshot).
• Can you commit to working Monday-Friday, 9am-5pm EST? Do you have experience we are looking for, including out tech stack?

Panchang Website Developer Needed (Swiss Ephemeris | No API)
Job Overview
We are looking for an experienced developer to build a Panchang (Hindu Calendar) website from scratch, using Swiss Ephemeris for accurate astronomical calculations.

Important: This project must be built without using any external Panchang APIs. All calculations should be done internally.

Project Goal
To develop a highly accurate, fast, and scalable Panchang system that generates daily Panchang details based on location and date.

Key Responsibilities

Panchang Calculation Engine
Implement Panchang logic using Swiss Ephemeris
Handle:
Latitude / Longitude
Time zones (IST + scalable for other regions)

Generate Daily Panchang Data
System should calculate and display:
Tithi
Nakshatra
Yoga
Karana
Sunrise & Sunset
Moonrise & Moonset
Rahu Kaal
Gulika Kaal
Yamaganda Kaal
Abhijit Muhurat

Calendar Interface
Monthly calendar view
Click any date → detailed Panchang
Highlight festivals & important days

Location-Based Panchang
Auto-detect user location OR manual selection
Support major Indian cities

Festival & Muhurat Logic
Display:
Ekadashi
Amavasya / Purnima
Major Hindu festivals
Based on Panchang rules (not static data)

Required Skills
Experience with Swiss Ephemeris integration (mandatory)
Strong knowledge of:
Astronomy-based calculations
Date/time & timezone handling
Backend experience (any):
Python / Node.js / PHP
Database:
MySQL / PostgreSQL

Deliverables
Fully functional Panchang calculation engine
Clean frontend UI
Optimized backend logic
Scalable codebase
Basic documentation
Restrictions
No use of Panchang APIs
No copied logic from other websites
Must be calculation-based system

Nice to Have
Multi-language support (Hindi / English)
Mobile responsive design
SEO-friendly structure

When Applying, Please Share:
Relevant past work (especially astrology/Panchang projects)
Experience with Swiss Ephemeris
Your approach to building this system
Estimated timeline
Budget quote

Project Type
Fixed price / Hourly (open to discussion)
Long-term opportunity for future enhancements

Skills: PHP, JavaScript, Python, MySQL, HTML, Node.js, PostgreSQL, Backend Development

We are seeking a skilled Power BI report editor to enhance our existing reports. The ideal candidate must be available for live collaboration via Microsoft Teams from 8 AM to 2 PM Central US time. The role involves making adjustments to reports based on team feedback and ensuring timely delivery of updates. Candidates should possess strong communication skills and a deep understanding of Power BI functionalities to contribute effectively to our projects.

• Weekend availability is a plus
• Describe your recent experience with similar projects
• What techniques would you use to clean a data set?
• How do you deal with outliers or missing values in a dataset?
• What tools do you use for data mining and visualization?

We are looking for an experienced developer or team to build a fintech remittance mobile application along with a simple professional website.

The app will allow users in Saudi Arabia to send money internationally to multiple countries worldwide through a simple, secure, and user-friendly system.

Important:
We will integrate with licensed remittance partners (API-based) for transfers, compliance, and settlement. We are not building our own remittance network.

---

Scope of Work:

1. Mobile App (User Side)
• User registration (Email + OTP)
• Add/manage recipients (bank or wallet details)
• Send money (amount, exchange rate, fee, receiver gets)
• Country selection (global support)
• Payment screen (card payment integration)
• Transfer tracking
• Transaction history

---

2. Payment Integration
• Integrate PayTabs / Stripe or similar
• Support Mada, Visa, MasterCard
• Handle payment flow and webhooks

---

3. Remittance Integration
• Integrate Nium or similar provider
• Support multiple countries (global transfers)
• Create transfers via API
• Fetch exchange rates
• Track transaction status

---

4. Backend System
• Scalable backend (Node.js preferred)
• Pricing engine (fee + FX margin)
• Quote system with expiry
• Connect payment → remittance flow
• Store transactions securely

---

5. Admin Dashboard
• View transactions
• Track transfer status
• Basic analytics
• User management

---

6. Website (Simple & Professional)

We also need a clean and attractive website for branding and customer trust.

Requirements:

• Landing page (professional design)
• About us section
• Customer support section
• Service information (global transfers)
• Promotional section (for future offers)
• Mobile-friendly design

---

Tech Preferences:
• Mobile: Flutter
• Backend: Node.js / NestJS
• Database: PostgreSQL

---

To Apply:
• Share similar fintech or payment projects
• Experience with API integrations
• Timeline and cost estimate

---

We are looking for a long-term developer. More features and countries will be added over time.

Thank you.

Title:
Senior Next.js Developer Needed to Finish Existing SaaS App (DeductIQ) + Launch Private Beta

Description:

I am hiring a skilled full-stack developer to continue building an existing SaaS platform called DeductIQ.

This is NOT a startup idea from scratch. I already have an existing codebase, product direction, domain, trademark filing submitted, and infrastructure accounts created.

The immediate goal is to launch a polished, stable private beta for 50 invited users for a 60-day testing period.

---

## About DeductIQ

DeductIQ is a modern SaaS platform that helps individuals, households, and small business owners organize finances and identify potential tax deductions throughout the year.

Core use cases include:

* Expense tracking
* Multi-business / multiple LLC organization
* AI-powered deduction guidance
* Document storage (receipts / tax files)
* Household + business financial organization
* Tax-readiness year-round

This is NOT tax filing software.

---

## Existing Stack

Current app already uses:

* Next.js
* Prisma ORM
* Neon Postgres
* Clerk Authentication
* Anthropic API
* Vercel hosting
* GitHub repository

There is already code written and UI established. I like the current design direction and want to continue from the existing foundation unless a specific rebuild recommendation is justified.

---

## Immediate Beta Goals

Need a developer who can take current codebase and get it truly beta-ready.

### Priority 1: Stabilize Existing Build

* Review current codebase
* Fix current Phase 2 errors / broken flows
* Resolve database / runtime issues
* Clean up unstable pages
* Ensure production deployment works on Vercel

### Priority 2: Invite-Only Private Beta Access

Implement secure beta gating:

* Only approved emails can register
* Non-approved users blocked
* Admin can manage beta invite list

### Priority 3: Complete Core User Experience

Must function smoothly:

* Sign up / login
* Guided onboarding flow
* Add household/business profiles
* Add expenses
* AI deduction scan results
* View/edit history
* Dashboard summary
* Clean navigation

### Priority 4: Document Vault

Need built-in secure file storage for:

* PDF
* JPG
* PNG

Features:

* Upload
* Preview
* Download
* Delete
* Link file to expense or business entity

### Priority 5: Beta Stability

Need polished MVP quality:

* Error handling
* Loading states
* Mobile responsive
* Fast performance
* No obvious broken pages
* Clean UX

---

## Branding / Design Standards

Current brand direction is modern, premium, clean, trustworthy.

### Colors

Primary Accent:
Gold / Green tone (final hex provided later, example #E8A020)

Background:
Dark Navy #07090F

### Typography

Headings:
Fraunces

Body:
Inter

### Feel

Need UI to feel:

* premium SaaS
* modern fintech
* clean and uncluttered
* trustworthy
* smooth onboarding like TurboTax / premium finance tools

Avoid generic template feel.

---

## What I Need From You

Please reply with:

1. Your experience with Next.js / Prisma / Clerk / SaaS apps
2. How soon you can start
3. Estimated time to beta-ready launch
4. Estimated cost range
5. Whether you would continue existing codebase or recommend partial rebuild
6. Links to previous projects

---

## Important

I am not looking for the cheapest developer.

I am looking for someone competent, efficient, honest, and able to help move quickly.

Strong communication matters.

---

## Future Work After Beta

If beta goes well, there will be paid follow-on work for:

* Billing / subscriptions
* Mileage tracker
* Tax calculators
* CPA portal
* Notifications
* Advanced AI tools

---

Please start your reply with the word “DeductIQ” so I know you read the full post.

Dog walker marketplace requirements for backend.

Background of the app:

The project is a marketplace application that connects dog owners (clients) with dog walkers (service providers), including an admin system for management. The frontend of the application is already completed, and I now need a backend implementation.

The system includes the following core features:

Full support of Hebrew language which is already implanted in the flutter file
User authentication (clients and walkers) using phone number and OTP (Firebase)
Booking system logic for dog walking and dog sitting services
Walker availability and service management
Real-time chat between users and walkers
Live booking state updates (walk/sit tracking) real-time GPS streaming
Cancellation logic with time-based rules and penalties
Reporting system for issues during active bookings
Admin visibility for reports and user activity manual via admin panel
Premium user features (e.g. booking flexibility, multiple dog profiles)
Local provider payment interagration

The backend must support a structured marketplace system with clear booking states, real-time updates, and secure data handling.

Backend architecture (schema + relationships)
Core booking system (all service types, controlled edge cases)
Basic cancellation logic (time-based, simplified penalties)
Payment integration (core flow)
Chat (real-time)
Admin panel (users + bookings + basic moderation)
APIs for your current frontend

Includes everything in MVP +

Complete booking state machine (all edge cases + transitions)
Advanced cancellation + penalty automation
Full payment lifecycle (refunds, payouts, edge scenarios)
Real-time architecture (tracking-ready, scalable events)
Advanced admin (reports, moderation system, controls)
Premium feature system

further documents attached here

Figma file: https://www.figma.com/design/AKh09ht91bqmsxytahjUAl/Doggy-project-update?node-id=0-1&t=Hs25RlBrOsxQ23zC-1

Source code will be sent once we will come to an agreement

We have a working RAG pipeline built on LangChain, FastAPI, and Supabase pgvector that has been in production for about two months. Recently, answer quality dropped noticeably — the model started pulling in irrelevant chunks and citing the wrong sections of documents even when the correct content is clearly in the store.

The pipeline is not broken. Queries return results and answers generate. The issue is retrieval precision — top-k chunks returned are frequently off-topic or from the wrong document. We believe the problem is somewhere in our chunking strategy, embedding model mismatch after a dependency bump, or a metadata filter we removed during a refactor.
Review our chunking config (chunk size, overlap, splitter type) against our document corpus and identify any obvious mismatch Confirm the embedding model used at ingestion time matches the one used at query time - we suspect a silent mismatch after upgrading langchain-openai Check our similarity search call - cosine vS. L2 distance config in pvector, and whether our index type (ivfflat vs. hnsw) is appropriate for our corpus size (~4,000 chunks)
Run 5-10 test queries against the live vector store and document which chunks
are being retrieved vs. which should be Deliver a short written diagnosis (no more than one page) with the root cause and the specific code change(s) needed to fix it

We are auditing infrastructure for the Southeast Asian market. We need a local developer to write a simple Node.js/Next.js script that tests query execution time (TTFB) between a standard AWS ap-southeast-1 (Singapore) database and a custom Postgres node we have deployed locally in Bangkok.

Deliverable: A short 1-page PDF of the test results.

We are looking for an experienced developer to create an AI-powered sales and customer service automation system. This system is for a company in Cartagena and will initially focus on WhatsApp integration.

The system must fulfill the following key functionalities:

Automated customer service via WhatsApp as the primary channel;
ability to efficiently answer frequently asked questions;
generation of personalized product quotes;
automatic lead classification (cold, warm, hot) to optimize follow-up;
closing sales outside of business hours;
registration and updating of customer information in Bitrix24 (CRM);
automated customer follow-up management;
and sending organized information to the design team for product customization.

The current technology stack, while flexible, includes:

n8n for automation orchestration;
OpenAI (GPT-4o) for AI capabilities;
PostgreSQL as the database;
ManyChat/WhatsApp API for communication;
and Bitrix24 as the customer relationship management (CRM) system.

We are looking for a professional with the following profile:

Demonstrable experience in n8n and building complex workflows.
Ability to integrate diverse APIs and systems.
Solid understanding of automation logic and scalable system design.
Capacity to propose technical improvements and innovative solutions.

Important project considerations:

The work will be carried out in phases, starting with the implementation of WhatsApp functionality.
There is the possibility of a long-term relationship for system maintenance and the development of new modules.
Delivery of basic system documentation is required.

When submitting your proposal, please include:

Details of similar projects you have completed.
The technology stack you are proficient in and propose for this project.
Your availability and estimated timelines for the initial phase.
Category IT & Programming
Subcategory Artificial Intelligence
Project size Medium

Delivery term: Not specified

Skills needed

API CRM JavaScript System Analysis Web Services Python

Skills: PHP, Java, Software Architecture, CRM, MySQL, PostgreSQL, API, Automation, OpenAI, n8n

Dynamic, experienced back-end developer needed for a custom graph-mapping web-app featuring intelligent back-end layers. Envision Obsidian reimagined with greater depth through graph-RAG, AI orchestration, and 3D functionality. The front-end has already launched. We are contracting to complete complex workflows and evolve them into highly dynamic back-end processes built around custom ontologies. United States residents only. Must have prior experience with RAG and visual knowledge systems. Must be able to sign an NDA.

Development phases: n8n ingestion pipelines → Python scripts → graph construction & querying → AI integration → 3D visualization layers.

Applicants must have experience with experimental projects, complex systems, and custom ontologies, and be willing to contribute suggestions for improving existing pipelines.

Skills needed short term:

-n8n – complex ingestion pipelines and scraping
-Python – asynchronous processing
-Neo4j Cypher
-Web development, API, and HTTP fundamentals
-Experience with local models (Ollama)
-Docker and containerization

Databases

-Postgres + PG Vector
-Neo4j (graph creation + advanced querying)
-Vector databases (Pinecone, Qdrant, or equivalent)

Skills needed long-term:

-3D layout, design, and animation
-React + front-end experience
-AWS infrastructure
-Semantic searches and language-to-query translation
-Hybrid retrieval techniques (vector similarity + graph traversal, e.g., vector + Cypher)
-Performance tuning and indexing (vector index optimization, graph query performance, scaling)
-Graph analytics: community detection, clustering (K-Means, hierarchical), grouping/partitioning logic, similarity measures, centrality, pathfinding
-Experience with highly complex, nonlinear workflows
-Vision LLM models and advanced OCR

Essential Requirements

-Demonstrated experience building experimental or custom ontologies
-United States residency and legal authorization to work required
-Willingness and ability to sign an NDA

Bonus Skills

-Node-based 3D systems (Maya, Blender, Houdini)
-Langflow
-Graphic design
-LLM training or foundational model creation

How to Apply

-Submit a proposal on Upwork that includes:
-Brief proof of experience with n8n pipelines or Neo4j/Cypher
At least one example of prior experimental graph or ontology work
-Your hourly and/or monthly rate

This is a bootstrapped startup seeking a contractor with strong potential for ongoing or full-time engagement.

LLM-generated proposals will likely be ignored.

• Please list any certifications related to this project
• Describe your recent experience with similar projects

Hello, I have a developed electron programme for my browser game. There are puzzles in game and I have a trained AI model but when puzzle showed up my programme couldn't solve I couldn't find problem every time it solves incorrectly. Do you have expertise in this area?

I am looking for an experienced developer to create a Telegram bot that manages access to two channels: one free (ad-supported) and one premium (ad-free). The bot will use the new Telegram Stars (XTR) system for payments and feature a referral system to incentivize growth.

Key features and functionality:

The core requirement is the full implementation of the Telegram Stars (XTR) payment flow, specifically for a single, one-time transaction that grants permanent access to the premium channel. The developer must handle invoices, pre-checkout queries, and successful payment confirmations via the Telegram Bot API. Once a payment is verified, the bot should automatically provide a unique invite link to the premium channel and log the user as a "lifetime member" in the database.

A referral system is required in order to track a users invites to the bot and automatically reward a user with access to the ad supported channel once 3 legitimate invites have been reached.

A secure Admin Menu is required to manage the business. This dashboard should display a leaderboard of top referrers and a detailed history of all Star payments. The admin should be able to see total revenue, view a list of all lifetime members, and have the ability to manually grant or revoke access if necessary. The admin menu can be integrated into the bot.

• Describe your recent experience with similar projects

1. Project Overview
ClearZone is a cleaning compliance SaaS platform built for cleaning contractors and their clients. It replaces paper checklists, manual timesheets, and phone calls with a single guided digital system — giving operatives clarity on the job and giving clients verifiable proof that the work was done.

This document is the full product brief for the custom build of ClearZone. It covers every module, every role, the complete data model, integration requirements, and commercial terms expected of the development engagement. It supersedes any earlier specification and should be treated as the single source of truth.

1.1 The Four Products
ClearZone is made up of four interconnected products:

Product User Purpose
Operative App (PWA) Cleaning operative Clock in, complete zone checklists, take photos, log chemicals, report issues, manage profile
Ops Dashboard Ops manager / supervisor Live site overview, scheduling, timesheet approval, alerts, reporting, staff management
Client Portal Facilities manager / client Read-only view of compliance data, zone status, photo evidence, SLA reports
Stock Management Ops manager + operative Track consumable stock levels, set par levels, receive low-stock alerts, raise purchase orders

1.2 What Already Exists
A working visual prototype has been built in Lovable (a no-code platform) to demonstrate the product concept. This prototype is for demonstration purposes only — it runs on hardcoded demo data with no backend. The custom build described in this brief starts from scratch using the technology stack defined in Section 3.

Key principle
The operative app is the foundation. Everything else — the dashboard, client portal, and agent intelligence layer — depends on the data the operative app collects. Build quality and reliability here is non-negotiable.

 
2. Gaps in the Existing Specification
The previously shared Operative App Technical Specification (Version 1.0, April 2026) covers the operative PWA in strong detail. However, the following areas are either absent or insufficiently specified and must be addressed in the build:

2.1 Ops Dashboard — Not Specified
The existing spec references a 'management console' briefly but does not specify it. The Ops Dashboard is a full web application requiring its own detailed design. Key screens required:
• Live site overview — all sites, real-time zone status per operative, alert feed
• Operative location tracking — based on last QR code scanned
• Timesheet review and approval workflow
• Zone and task template builder — create/edit zones, assign task checklists per zone
• QR code generation and print-ready PDF export
• Staff scheduling and shift assignment
• Direct messaging to operatives and bulletin broadcast
• Chemical library and COSHH management
• Credential and training management — verify uploads, track expiry, set site requirements
• Site induction builder and completion tracking
• Reporting — SLA performance, zone completion rates, issue resolution, compliance summary

2.2 Client Portal — Not Specified
The spec mentions the client portal only as a Phase 3 item without any detail. The Client Portal is a separate read-only web interface for the end client (facilities manager). Required screens:
• Live zone status view — which zones are complete, in progress, or outstanding
• Photo evidence feed — timestamped and geotagged completion photos per zone
• SLA performance dashboard — daily, weekly, monthly compliance metrics
• Issue report log — all flagged issues with status and resolution notes
• Daily PDF report access — downloadable compliance reports per site
• Notifications — alerts when SLA thresholds are breached

2.3 Stock Management Module — Not Specified
The existing spec does not mention stock management. This is a required module, detailed fully in Section 5 of this brief.

2.4 Authentication and Role-Based Access — Underspecified
The spec does not define the authentication system or role hierarchy. This must be built explicitly:
• Authentication via Supabase Auth (email/password + magic link)
• Four distinct roles: Operative, Supervisor, Ops Manager, Client
• Row-level security enforced at the database level — not just frontend routing
• Operatives can only see their own shifts, assigned zones, and their own profile
• Supervisors can see all operatives at their assigned sites
• Ops Managers have full access across all sites in their company account
• Clients have read-only access scoped to their contracted sites only

2.5 Multi-Tenancy — Not Specified
ClearZone is a SaaS product serving multiple cleaning companies. The data model must support multi-tenancy from day one:
• Each cleaning company is a separate 'Organisation' with isolated data
• Operatives, sites, zones, and clients belong to an Organisation
• No data leakage between organisations at any layer
• Organisation-level settings: branding, language defaults, notification preferences

2.6 Pricing and Subscription — Not Specified
ClearZone has three pricing tiers (Starter £49/site/month, Professional £89/site/month, Enterprise £149/site/month). The system must enforce feature gating based on the active subscription tier:
• Starter: up to 3 sites, up to 20 zones per site, core app + dashboard + daily report email, CSV payroll export
• Professional: unlimited sites and zones, client portal, 10+ languages, UK payroll integrations, lone worker safety, SLA reporting
• Enterprise: white-label branding, multi-client management, advanced analytics, offline mode, dedicated account manager, API access
• Subscription managed via Stripe — billing, upgrades, downgrades, and cancellation

2.7 Payroll Integrations — Underspecified
The spec lists payroll integrations as a Phase 3 item but does not define the scope. The following integrations must be specified and built:
• Xero — approved timesheet data exported as draft invoices or payroll items
• Sage Payroll — CSV or API export in Sage-compatible format
• BrightPay — CSV export in BrightPay import format
• Generic CSV export — fallback for all other systems
• Each export must handle: operative name, site, clock-in/out times, total hours, break deductions, pay rate (if configured), approval status

2.8 IP Ownership and Code Repository — Not Specified
The following must be agreed before work commences:
• All intellectual property created in this engagement belongs solely to ClearZone Ltd
• The codebase must be hosted in a GitHub repository owned by ClearZone Ltd from day one
• The developer must not retain any copy of the codebase on project conclusion
• All third-party libraries used must be permissively licensed (MIT, Apache 2.0, or equivalent)

 
3. Technology Stack
The following stack is specified in the existing operative app spec and is confirmed as the target stack for all ClearZone products:

Component Technology Applies To Notes
Frontend framework Next.js (React) + TypeScript All products Single monorepo preferred
Styling Tailwind CSS All products Utility-first, responsive by default
Operative offline storage IndexedDB via Dexie.js Operative PWA Full offline shift capability required
Service worker / sync Workbox Operative PWA Background sync queue for offline data
QR scanning html5-qrcode Operative PWA Browser-native, no plugins
Push notifications Web Push API + Firebase Cloud Messaging Operative PWA Works when app is not open
State management Zustand All products Persisted to IndexedDB on operative
Internationalisation next-intl Operative PWA 10+ languages at launch
Database PostgreSQL via Prisma ORM Backend With row-level security via Supabase
Authentication Supabase Auth All products Email/password + magic link
File storage Cloudflare R2 or AWS S3 All products Photos, documents, PDFs
Backend hosting Railway or Vercel Backend + frontend Auto-deploy from GitHub
Payments Stripe SaaS billing Subscriptions, upgrades, cancellation
Email Resend or SendGrid All products Daily reports, notifications, alerts

 
4. Operative App (PWA)
The operative app is a Progressive Web App that runs on any device with a browser. It is the primary data collection tool. All design decisions prioritise accessibility, offline capability, and simplicity for non-technical, multilingual users.

4.1 Core User Journey
• Arrive at site → Clock In (GPS geofence or QR fallback)
• View assigned zones for the shift
• Scan QR code at each zone to start work
• Complete task checklist (mandatory tasks enforced)
• Log chemicals used
• Take mandatory completion photo (timestamped, geotagged)
• Flag any issues with photo and description
• Move to next zone
• Clock Out at end of shift — summary shown, timesheet auto-submitted

4.2 Screens Required
Screen Key Elements Notes
Clock In Large button, GPS check, QR fallback, green tick animation on success Vibration feedback on mobile
Zone List Zones in order, colour status (grey/blue/green), progress bar Pull-to-refresh, offline works identically
Zone Active (Checklist) QR scan to enter, task list with large tick boxes, chemical log, photo capture, issue flag button Cannot skip required tasks
Messages Direct messages + Bulletins tabs, text/voice send, push notifications Offline queue for outbound messages
My Hours Read-only timesheet — shifts, hours, approval status No editing capability
My Profile Personal details, document upload (ID, DBS, certs), expiry alerts, language selector Guided first-time setup flow
Clock Out Summary (zones done, issues flagged, total hours), confirm button GPS or QR location check
Stock (Operative) Simple list of products at this site, stepper to update current volume Read/update only — no order management

4.3 Languages at Launch
Minimum 10 languages: English, Polish, Romanian, Portuguese, Spanish, Bengali, Urdu, Hindi, Arabic, Somali. Each language is a JSON translation file — additional languages must be addable without code changes.

4.4 Offline Capability
Hard requirement
The operative app must function with zero internet connectivity for an entire shift. This is a hard requirement, not a nice-to-have.

• On last connectivity before shift: app downloads all zones, checklists, chemical lists, and language pack for the day
• All operative actions (clock-in, QR scans, task ticks, photos, issue flags) write to IndexedDB
• Background sync queue processes in order when connectivity returns
• Retry with exponential backoff on sync failure
• Photos queue behind text data and upload progressively
• Operative sees green/grey connectivity indicator only — no error messages, no functionality difference

4.5 Lone Worker Safety
• Activates automatically when only one operative is clocked in at a site
• Periodic check-in prompt every 15 minutes (configurable) — single tap to confirm
• Missed check-in triggers immediate alert to supervisor with last known location
• Escalation to secondary contact if supervisor does not acknowledge within 5 minutes

 
5. Ops Dashboard
The Ops Dashboard is a web application for cleaning company operators and site managers. It is the control centre that sits above the operative app — taking the data operatives generate and turning it into actionable management information.

5.1 Screens Required
Screen Key Elements Notes
Home / Overview All active sites, zone completion %, live operative count, alert feed, overdue zones highlighted Real-time updates via websocket or polling
Site Detail Zone-by-zone status for one site, operative locations (last QR), time since last activity Drill down from Overview
Operative Management All operatives list, status (active/inactive), credential status, site assignments Search and filter
Scheduling Assign operatives to sites and shifts, set zone order, handle ad-hoc requests Dispatch urgent task to nearest available operative
Timesheet Review All pending timesheets, approve/query per shift, flag discrepancies Approved timesheets feed payroll export
Zone Builder Create/edit zones, assign tasks, set required vs optional, set chemical list per zone, generate QR codes Print-ready PDF of all QR codes per site
Task Library Global library of task templates reusable across zones
Chemical Library COSHH-compliant product list with hazard symbols, approved list per site
Staff Profiles View operative profile, verify credential uploads, set required credentials per site, training expiry alerts Manual upload capability for in-house training records
Site Inductions Build induction content per site, track completion by operative, block clock-in until complete
Messaging All operative message threads, bulletin creation and broadcast, read receipts Urgent flag for safety alerts
Reports SLA performance, zone completion rates, issue resolution times, chemical usage, training compliance Export to PDF and CSV
Settings Organisation profile, site management, subscription and billing, notification preferences, language defaults

5.2 Payroll Export
From the Timesheet Review screen, approved timesheets can be exported in the following formats:
• Xero — API push or CSV in Xero payroll import format
• Sage Payroll — CSV in Sage-compatible format
• BrightPay — CSV in BrightPay import format
• Generic CSV — configurable column mapping for all other systems

5.3 Daily Report Email
Each morning, an automated PDF report is generated per site and emailed to the nominated client contact. The report must include:
• All zones cleaned the previous day — time started, time completed, operative name
• Completion photo for each zone
• Chemical log per zone
• Any issues flagged with photos and current status
• SLA compliance summary (percentage of zones completed on schedule)
• Hours summary per operative
The report is branded with the cleaning company logo (Professional and Enterprise tiers). It is sent automatically — no manual action required.

 
6. Client Portal
The Client Portal is a separate, read-only web interface for facilities managers and building clients. It gives clients transparent, real-time visibility of cleaning compliance at their sites without access to operational or HR data.

6.1 Access
• Clients access via a unique login (email + magic link — no password to manage)
• Each client sees only the sites they are contracted for
• No access to operative personal details, payroll data, or internal management information

6.2 Screens Required
Screen Key Elements Notes
Site Dashboard Live zone status for each site — complete, in progress, outstanding. Operative count currently on site. Refreshes in near real-time
Zone Detail All tasks completed for a zone, completion photo, chemical log, operative name, timestamps Photo lightbox view
Issue Log All issues flagged at this client's sites, with photos, status, and resolution notes Filter by site, status, date
SLA Reports Daily, weekly, monthly compliance metrics — percentage complete, average completion time, trend charts Export to PDF
Daily Reports Archive of all daily PDF reports for this client's sites — downloadable
Notifications Email alert when SLA threshold is breached or a critical issue is flagged Configurable threshold per client

 
7. Stock Management Module
The Stock Management module tracks cleaning consumables (toilet rolls, hand soap, bin bags, cleaning fluid, etc.) per site. It alerts when stock falls below acceptable levels and enables purchase orders to be raised directly within the platform.

7.1 Data Model
Entity Key Fields Notes
Product id, name, category, unit (rolls/litres/bags/each), description Global product library, managed by Ops Manager
SiteStock siteId, productId, currentVolume, parLevel, lastUpdatedBy, lastUpdatedAt One record per product per site
StockOrder id, siteId, productId, quantityOrdered, status, raisedBy, raisedAt, deliveredAt, notes Status: Draft → Submitted → Ordered → Delivered
StockAlert id, siteId, productId, alertType (Low/Critical), triggeredAt, resolvedAt, resolvedBy Auto-generated when volume drops below par

7.2 Alert Logic
• Low alert: current volume falls below par level — item flagged 🟡 Low
• Critical alert: current volume falls below 25% of par level — item flagged 🔴 Critical
• Alert displayed in Ops Dashboard alert feed and stock overview
• Email notification sent to Ops Manager on Critical alert
• Alert resolves automatically when operative updates volume above par level

7.3 Ops Manager Screens
Screen Key Elements Notes
Stock Overview (/stock) All products across all sites, current volume, par level, status badge (OK/Low/Critical), last updated, Raise Order button per low item Filter by site, category, status
Site Stock Detail (/stock/[site-id]) Site-specific product list, inline editable current volume, visual progress bar (current vs par), last updated by/when Audit trail of all volume updates
Order Management (/stock/orders) All orders with status, filter by site/status/date, mark as Ordered/Delivered Auto-calculates quantity needed (par − current)
Product Library (/stock/products) Create/edit products, set categories and units Ops Manager only
Par Level Settings Set and edit par levels per product per site Ops Manager only — operatives cannot change par levels

7.4 Operative Stock Screen
• Simple scrollable list of products assigned to operative's current site
• Current volume shown with status colour indicator
• Stepper control (+/-) to update current volume after restocking
• Confirmation tap saves update with timestamp and operative name
• No access to par level settings, order management, or product library

7.5 Role Permissions
Feature Operative Supervisor Ops Manager
View stock levels ✓ ✓ ✓
Update current volume ✓ (own site) ✓ ✓
Set / edit par levels ✗ ✗ ✓
Create / edit products ✗ ✗ ✓
Raise orders ✗ ✓ (view only) ✓
Manage / close orders ✗ ✗ ✓
View order history ✗ ✓ ✓

 
8. Complete Data Model
All entities below are scoped to an Organisation (multi-tenancy enforced via row-level security in Supabase/PostgreSQL).

Entity Key Fields Relationships
Organisation id, name, logo_url, subscription_tier, stripe_customer_id, created_at Has many Sites, Users, Products
User id, org_id, name, email, role (operative/supervisor/ops_manager/client), language, photo_url, status Belongs to Organisation
Site id, org_id, name, address, client_id, timezone, active Has many Zones, SiteStocks, Shifts
Zone id, site_id, name, floor, order_index, qr_code_uuid, estimated_minutes Has many Tasks, ZoneVisits
Task id, zone_id, description, required (bool), order_index, translations (JSON) Belongs to Zone
Chemical id, org_id, name, coshh_category, hazard_symbols (array), approved_sites (array) Belongs to Organisation
Shift id, site_id, operative_id, clock_in_at, clock_out_at, clock_in_method (GPS/QR), gps_coords, status, approved_by, approved_at Has many ZoneVisits
ZoneVisit id, shift_id, zone_id, qr_scanned_at, started_at, completed_at, photo_url, photo_gps Has many TaskCompletions, ChemicalLogs
TaskCompletion id, zone_visit_id, task_id, completed_at, skipped (bool) Belongs to ZoneVisit
ChemicalLog id, zone_visit_id, chemical_id, quantity (optional) Belongs to ZoneVisit
IssueReport id, zone_visit_id, site_id, description, photo_url, priority, status, resolved_at, resolved_by Belongs to ZoneVisit
Message id, operative_id, manager_id, content, type (text/voice/photo), sent_at, read_at Direct message thread
Bulletin id, org_id, site_ids (array or null for all), title, content, urgency, created_by, created_at Broadcast to operatives
OperativeProfile id, user_id, emergency_contact, preferred_language, notes 1:1 with User
Credential id, user_id, type, document_url, issue_date, expiry_date, status (pending/verified/expired) Belongs to User
TrainingRecord id, user_id, training_type, provider, completed_at, expiry_date Belongs to User
SiteInduction id, user_id, site_id, completed_at, confirmed (bool) Per operative per site
SyncQueue id, user_id, entity_type, entity_id, payload (JSON), retry_count, last_attempt_at Operative-side offline queue
Product (Stock) id, org_id, name, category, unit, description Belongs to Organisation
SiteStock id, site_id, product_id, current_volume, par_level, last_updated_by, last_updated_at Per product per site
StockOrder id, site_id, product_id, quantity_ordered, status, raised_by, raised_at, delivered_at, notes Order lifecycle
StockAlert id, site_id, product_id, alert_type, triggered_at, resolved_at, resolved_by Auto-generated

 
9. Build Phases & Priorities
The following phased plan reflects the priority order for the build. Phase 1 must be complete and tested before Phase 2 begins. The operative app is the foundation — all other products depend on the data it generates.

Phase 1 — Operative PWA (Week 1)
The complete operative-facing application including:
• Authentication (Supabase Auth) — email/password + magic link
• Clock in / out with GPS geofencing and QR code fallback
• Zone list with status indicators and progress tracking
• QR zone scanning via device camera
• Task checklist with mandatory completion enforcement and order locking
• Photo capture with automatic timestamp and geotag
• Chemical logging from COSHH-approved list
• Issue flagging with photo, text and optional voice description
• Full offline capability with background sync queue and retry logic
• Multilingual interface — minimum 10 languages
• Push notifications (Web Push API via service worker)
• Messaging — direct messages and bulletin feed
• My Hours — read-only timesheet view
• My Profile — document upload, credential status, language selector
• Stock update screen — volume stepper, operative view only
• Lone worker check-in system
• PWA manifest and add-to-home-screen prompt

Phase 2 — Ops Dashboard (Week 2)
• Authentication and role-based routing
• Live site overview with real-time zone status
• Operative tracking (last QR scan location)
• Alert feed — overdue zones, missed clock-ins, flagged issues, lone worker alerts
• Timesheet review and approval
• Zone and task template builder
• QR code generation and print-ready PDF export
• Staff scheduling and shift assignment
• Direct messaging and bulletin broadcast with read receipts
• Chemical library and COSHH management
• Credential verification and training expiry tracking
• Site induction builder and completion tracking
• Stock management — full Ops Manager view (all screens from Section 7)
• Payroll export — Xero, Sage, BrightPay, CSV
• Daily PDF report generation and automated email delivery

Phase 3 — Client Portal (Week 3)
• Separate client-facing authentication (magic link)
• Live zone status and operative count per site
• Photo evidence feed per zone
• Issue log with status tracking
• SLA performance dashboard with export
• Daily report archive and download
• Email notifications on SLA breach or critical issue

Phase 4 — Integrations and Polish (Week 3–4)
• Stripe subscription integration — tier enforcement, billing, upgrades
• Xero, Sage, BrightPay API integrations (deeper than CSV where APIs are available)
• White-label branding for Enterprise tier — custom logo, colour scheme
• Performance optimisation for low-end devices
• End-to-end testing across device types (Android, iOS, old and new hardware)
• QR code print workflow and label formatting
• Onboarding flow for new organisations and first-time operatives
• Deploy to first live pilot site

 
10. Commercial Terms Required
The following terms must be agreed and documented in a signed contract before development commences. This section is not a legal document — it is a list of requirements that must be reflected in the formal agreement.

10.1 IP and Code Ownership
• All intellectual property developed during this engagement belongs solely to ClearZone Ltd from the moment it is created
• The GitHub repository must be created under and remain under a ClearZone Ltd-owned account
• The developer retains no rights to use, resell, or reference the codebase or product
• All third-party libraries must be permissively licensed (MIT, Apache 2.0, ISC or equivalent)

10.2 Pricing and Payment
• Quote must be broken down by phase — a single lump sum is not acceptable
• Fixed price per phase preferred; if time-and-materials, a cap must be agreed per phase
• Payment schedule tied to phase delivery milestones, not calendar dates
• No payment in advance for undelivered phases

10.3 Timeline and Overruns
• Start date, phase end dates, and final delivery date must be agreed in writing
• What happens if a phase overruns must be defined — does cost increase? Does timeline extend?
• ClearZone must have the right to pause or terminate at the end of any phase

10.4 Handover and Documentation
• On completion, all code must be documented to a standard that allows another developer to continue work
• All environment variables, third-party API keys, and infrastructure access must be handed over
• A deployment runbook must be provided — how to deploy, how to roll back, how to scale
• A 30-day post-launch support period must be included for bug fixes arising from the build (not new features)

I am looking for an experienced full-stack developer with strong AWS and AI integration skills to build a scalable, high-performance application.

Project Overview:
The goal is to develop a modern, secure, and user-friendly system with cloud deployment and AI-powered features.

Scope of Work:
- Full-stack application development (frontend + backend)
- Clean and professional UI/UX (mobile + desktop responsive)
- Backend development using Java (Spring Boot) / Node.js
- Secure authentication & authorization system
- Proper database design (MySQL / PostgreSQL)

AWS / Cloud Requirements:
- EC2 and RDS setup with start/stop automation
- Secure VPC configuration
- HTTPS setup (SSL)
- Deployment and performance optimization
- Cost monitoring and optimization

AI Integration:
- Integration of AI APIs (OpenAI or similar)
- Chatbot / smart automation features
- Efficient data handling and response optimization

Additional Expectations:
- Clean, scalable, and well-documented code
- Regular progress updates
- Ability to suggest improvements and best practices

Requirements:
- Proven experience in full-stack development
- Strong knowledge of AWS services
- Experience with AI integrations

This project has strong potential for long-term collaboration if the work quality is excellent.

Skills: Java, Website Design, MySQL, Amazon Web Services, Software Development, Full Stack Development, Automation, Chatbot, Spring Boot

Overview
We are building a real-world telemetry system for an Sub Polar expedition (Outer Edge Polar Challenge), where a wind-powered vehicle will operate continuously in sub-zero conditions with intermittent connectivity.

We are looking for a practical, execution-focused backend engineer who can:

Deploy a simple, reliable backend on Render
Handle telemetry ingestion and storage
Design and run a clean, real-time dashboard on a rugged tablet used in the field

This is not a theoretical project. It needs to just work under constraint.

Core Scope

1. Backend (Render)

Deploy API (Node.js or Python) on Render
Build telemetry ingestion endpoint (POST /telemetry)
Store data in PostgreSQL
Handle intermittent connectivity (buffered uploads, no data loss)
Create simple endpoints:
GET /latest
GET /history

2. Data Handling

Continuous low-volume telemetry (GPS, speed, wind, system state)
Design for:
burst uploads after reconnect
data integrity over latency
simple validation and logging

3. Dashboard (Critical)

Design a clean, minimal dashboard UI that runs on a rugged tablet in the field
Must:
display live telemetry (speed, position, key system metrics)
handle intermittent connection gracefully
auto-refresh and remain stable over long periods
be readable in harsh conditions (glare, low temps, simple UI)
Tech options open (React, simple web app, etc.)
Must run via browser or lightweight local interface

4. System Constraints

Intermittent Starlink connectivity
No live debugging during operation
System must degrade gracefully, not fail
Simplicity and reliability

We're a US-based SaaS company with a mature Ruby on Rails application (Rails 7, PostgreSQL, Sidekiq, Redis) serving B2B customers. We're looking for a senior Rails engineer to join as a long-term partner for ongoing product work, with a strong focus on integrating AI features into our existing stack.

About the work:
- Core Rails product development: new features, refactors, performance work, and background processing
- AI integration: OpenAI / Claude API work, RAG pipelines, embeddings, vector search, streaming responses, prompt engineering in production
- Multi-tenant architecture, API integrations with third-party SaaS tools, and robust background jobs
- Code review, architecture input, and occasionally pairing with our in-house team
- Production-grade work: you care about tests, observability, and not breaking things that already work

What we're looking for:
- 8+ years of professional Ruby on Rails experience, ideally with multiple production Rails apps you've owned end-to-end
- Strong PostgreSQL skills (query tuning, indexes, schema design)
- Real LLM integration experience — not just demos. You've shipped OpenAI or Claude features that real users depend on
- Comfortable with Sidekiq, background jobs, and high-throughput workflows
- Clear written communication in English, comfortable with async work across time zones
- Top Rated / Top Rated Plus freelancers preferred; 100% Job Success Score strongly preferred

Engagement details:
- Long-term, ongoing — we're looking for someone to stick with us for 12+ months, not a one-off project
- Hourly, roughly 20-30 hours per week to start, with flexibility to scale up
- Budget: $40-$75/hr depending on experience and AI specialization
- You can work from anywhere, but we need 3-4 hours of overlap with US Eastern time

When you apply, please share:
1. One or two examples of production Rails + AI work you've shipped, and what your specific contribution was
2. A short note on how you'd approach adding an LLM-backed feature to an existing Rails app without making it fragile
3. Your current availability

We'll shortlist quickly and start with a small paid trial task before any long-term commitment.

Summary:

We are looking for a Senior Backend Engineer (Node.js & TypeScript) to join a small, technical team building an AI-powered inventory and demand planning platform for Shopify brands. The role focuses on designing and scaling backend systems that power data synchronization, forecasting, and integrations with external platforms. You will work closely with the founder and engineering team to deliver reliable, high-impact systems that directly influence business decisions for real customers.

Responsibilities:
- to own and improve the core data sync architecture (Shopify API, webhooks, job queues);
- to build and maintain forecasting and demand planning systems;
- to design clean, scalable data models for multi-store and multi-location inventory;
- to develop and maintain integrations with third-party platforms (Amazon, 3PLs, WMS);
- to collaborate directly with the founder on architecture and technical decisions;
- to ensure system reliability, performance, and correctness in production environments;
- to participate in code reviews and maintain high engineering standards.

Core Qualifications:
- at least 5+ years of experience as a Backend Engineer;
- strong proficiency in Node.js and TypeScript;
- solid experience with PostgreSQL and data modeling;
- experience working with APIs, webhooks, and distributed systems;
- understanding of asynchronous processing (job queues, event-driven architecture);
- experience with cloud environments (AWS preferred);
- strong problem-solving skills and attention to detail;
- ability to work in a small, fast-moving team.

Nice to Have:
- experience with Shopify ecosystem and integrations;
- experience with Kafka or similar event streaming systems;
- exposure to Kubernetes and CI/CD pipelines;
- experience with forecasting, supply chain, or inventory systems;
- previous experience in startups or product-driven environments.

Please note:
* Strong English communication skills are required.
* This is a contractor position; no employee benefits are included in the contract.