AI Code Generator Projects on Upwork

*USA EST NEW YORK TIME ZONE - 830 AM TO 5 PM EST *
MESSAGE ONLY IN UPWORK, NOT OUTSIDE, WILL BE REPORTED - THANK YOU!

AutoBuffy.com is a fast-growing US online auto parts store with over 1 million parts, built on Next.js 14 (App Router), FastAPI, MySQL, Typesense and Stripe.

We're hiring one senior engineer to own conversion: everything from the moment a shopper lands to the moment they pay. This isn't a "close tickets" role. You own the funnel, and you're measured on how many visitors turn into orders.

What you'll own

1. Speed. Make the site fast on real phones, not just in lab tests. Targets: product page under 2.5 s to show its main content on mobile, homepage under 1 s, and no change that makes a page slower.
2. Search and filters. Shoppers find the right part on the first try: typed searches, part numbers, year/make/model, and filters that work. No dead ends and no zero-result searches for common parts.
3. Product pages. Make "will this fit my car?" obvious. Price, delivery date and Add to Cart clear above the fold on mobile. Trust signals that are real, never fake.
4. Mobile cart, checkout and chat. Fewer steps, Apple Pay and Google Pay, help in the moment, and fewer people dropping off.
2. Search and filters. Shoppers find the right part on the first try: typed searches, part numbers, year/make/model, and filters that work. No dead ends and no zero-result searches for common parts.
3. Product pages. Make "will this fit my car?" obvious. Price, delivery date and Add to Cart clear above the fold on mobile. Trust signals that are real, never fake.
4. Mobile cart, checkout and chat. Fewer steps, Apple Pay and Google Pay, help in the moment, and fewer people dropping off.
5. The new homepage. Finish and launch our new search-first homepage (a design and working prototype exist).
6. Measurement. Set up and own the funnel numbers: visit, then search, then product page, then cart, then checkout, then order. Every change ships with before and after data.

How we work

- Small pull requests, CI must pass, staging first, tested on desktop and a real phone before launch.
- Every proposal says how Amazon, AutoZone, RockAuto or carparts.com handle the same thing, and where we fall short.
- You'll work with our platform engineer (backend and catalog) and report to the CEO.
- Hours overlap with US Eastern time, at least 4 hours a day.

First 90 days, what success looks like

- Mobile product pages under 2.5 s, measured from real-user data.
- Visitor-to-order conversion up 15% or more.
- Search-to-purchase rate up, zero-result searches down.

Goal is to own the store front:
Homepage
Search
Product Page
Cart
Checkout
User Dashboard
Internal SEO/AIO Pages

Experience Level: Intermediate to Expert
Project Type: One time project, with potential for ongoing work
Budget: $10-$15 fixed price, depending on the final scope
Job Description
I'm looking for an experienced AI Agent Engineer to help me build a practical AI assistant that can handle selected tasks, connect with external tools, and automate repetitive workflows.
I want to explore a solution using tools such as OpenClaw, Hermes Agent, MCP, and n8n. My goal is to create a reliable system that goes beyond a basic chatbot and can execute defined tasks through approved integrations.
Scope of Work
Recommend a suitable AI agent architecture based on my requirements.
Configure an AI agent and connect it to selected tools using MCP or supported integrations.
Build n8n workflows to automate specific tasks.
Set up secure access, permissions, and error handling.
Test the system and provide basic documentation for future maintenance.
Ideal Candidate
Hands-on experience building AI agents and tool-calling workflows.
Familiarity with OpenClaw, Hermes Agent, MCP servers, or similar frameworks.
Experience with n8n, APIs, webhooks, and automation.
Understanding of security, access controls, and reliable workflow execution.
Please share examples of similar projects you have built and briefly explain how you would approach this task.
I'd prefer to start with a clearly defined, working prototype and expand it after the initial results are satisfactory.

I'ld like to generate ai gents for my whatsapp marketing and calling the leads which are generated.

i want to connect dealmachine, a website that gives me real estate leads (proeprties and their informations) i want to see if i can connect that to my command center in base 44

We are seeking a skilled software developer familiar with Claude Code/Cursor to join our team. The ideal candidate will have extensive experience in software engineering, capable of reviewing, debugging, and maintaining code efficiently. Your role will involve utilizing Claude Code/Cursor as a development accelerator while ensuring high-quality code standards. Proven collaboration skills and a proactive approach to problem-solving are essential for success in this position.

About the Role
We are a digital marketing agency for nonprofits and mission-driven organizations, and we also build our own software for the nonprofit sector. We are looking for a part-time developer who lives in Claude Code and understands digital marketing from the inside, to turn working prototypes into tools people actually use.

This is a contract role at 10 to 20 hours a week, around $30 an hour, fully remote. You will work alongside a strategist who scopes each project and owns the client relationship, and you will handle the bulk of the build.

What You'll Work On
Your time splits between client projects for nonprofits and internal tools for our own team and software product. A strategist typically builds a quick proof of concept in a few hours; you take it from there to something polished, reliable and ready to hand over.

Client builds. Custom internal tools for nonprofits: dashboards and insights on top of their CRM and fundraising data, custom databases, and automations that replace manual spreadsheet work.

Campaign tooling. Finishing and extending a campaign builder that creates Google and Meta ad campaigns through their APIs, including targeting options like custom and lookalike audiences.

Reporting automation. Piping ad platform and analytics data into Google Sheets and dashboards through tools like Windsor.ai, and drafting performance reports from live data.

Optimization audits. Pulling account data via API and checking it against our internal best-practice docs, so account managers get a regular list of recommended changes to review.

Research and data tools. Scrapers, enrichment and analysis across large datasets of nonprofit websites, including search and AI visibility tooling.

Documentation. Writing clear setup and maintenance notes so clients and teammates can run and extend what you build.

What We're Looking For
Deep, daily experience with Claude Code. You use it as your main development environment, and you know how to set up projects, write instructions and skills, connect MCP servers and APIs, and manage token usage across models.

A background in digital marketing. You have worked hands-on with Google Ads, Meta Ads, GA4 or similar platforms, so you understand how campaigns, audiences and conversion tracking work without needing them explained.

A portfolio of things you have built. Apps, automations, dashboards or agents you can show us, with a short note on what each one does and how you built it.

Comfort with APIs and integrations. You can read API documentation, handle authentication, and connect tools like Google Sheets, CRMs and ad platforms, including through middleware such as Composio or Zapier.

Judgment about what to build. You can tell the difference between a demo and a tool someone will use every week, and you build for the second.

Clear written communication. You keep a strategist up to date asynchronously, flag blockers early, and document your work so someone else can pick it up.

A human-in-the-loop mindset. You build tools that recommend and draft, with a person approving any change to a live system.

Nice to Have
Experience writing code by hand. A background in traditional development (Python, JavaScript or TypeScript, SQL) before AI tools, so you can read, debug and simplify what the AI writes rather than just accept it.

Experience with the Google Ads and Meta Marketing APIs specifically.

Experience with CRMs common in the nonprofit sector, such as Salesforce, Bloomerang or HubSpot.
Familiarity with databases and hosting, such as Supabase, Postgres, Airtable or Vercel.

Experience with open-source or self-hosted models, or tools that route tasks across models to keep costs down.

Experience working with nonprofits or mission-driven organizations.

Comfort joining the occasional client call to demo a tool or gather feedback.

How We Work
We are a small, remote team that leans hard on AI and holds a high bar for quality. You will report to a strategist who scopes each project with the client, reviews your work before it ships, and stays accountable for delivery. Our founder sets the product direction and stays close to the build.
Most communication is asynchronous in Slack, with short Loom walkthroughs instead of long meetings.
Projects arrive with a clear scope and often a working prototype, so you can start building on day one.

You will have access to our shared Claude Code setup, internal skills and best-practice docs, and the connected accounts each project needs.
We want to learn from you too. If you know a better way to build something, we want to hear it.

How to Apply
Send us:
Links to two or three things you have built with Claude Code, with a sentence on each about the problem it solved.
A short note on your digital marketing experience: which platforms, and what you did in them.
If you have written code by hand before, tell us what languages and what you built.
Your weekly availability and hourly rate.
Shortlisted candidates will go through two rounds of interviews:
An interview with our lead strategist.
An interview with our founder and CEO/COO.
The process may also include a short paid test build.

I am currently reviewing developers for the Beta development of Core16 Personalities, a personality assessment and AI-supported personal development platform.



A substantial amount of the product architecture has already been completed before development begins. I have developed a structured set of specifications covering the assessment architecture, scoring, interpretation, results, reporting, AI boundaries, data structures, governance and wider product requirements.



The project is intended to be built using a Lovable-first / AI-assisted development approach.



My objective is not to commission a conventional development project in which every component is manually coded from scratch where Lovable or other appropriate AI-assisted tools can reasonably produce it.



Instead, I am looking for a developer who can use the existing specifications to:



- translate the architecture into an effective Lovable/AI build strategy;
- create appropriate structured prompts and implementation instructions;
- allow Lovable/AI to perform as much of the implementation as reasonably possible;
- provide specialist development where custom work is genuinely required;
- review and correct AI-generated implementation;
- handle integration, security and technically sensitive elements;
- test and debug the system; and
- take responsibility for ensuring that the agreed Beta works properly from end to end.



For the initial review, I can provide the Core16 Steps 1–19, which describe the implementation architecture in considerable detail.



There is also a separate underlying documentation set, D01–D23, covering the wider product, scientific, governance, AI and architectural foundations. If, after reviewing summaries, you believe particular documents are required, please identify them and I can provide the relevant material.



What I would like from you



Before development begins, I would like you to review the material and recommend:



1. What should constitute a coherent, functioning Core16 Beta suitable for controlled real-world users.



2. What functionality should deliberately be deferred to the Full Version/Post-Beta.



3. Which elements can reasonably be produced using Lovable/AI.



4. Which elements genuinely require specialist developer intervention.



5. What you would expect me, as the founder/product owner, to handle directly using Lovable/AI.



6. What you would take responsibility for yourself.



7. Any technical dependencies, risks or missing specifications you identify.



I would then like us to agree a Beta vs Full Version Scope Matrix and objective Beta acceptance criteria before development starts.



Commercial structure



My preference is to commission the agreed Beta as a fixed-scope, fixed-price project, rather than purchasing an open-ended allocation of development hours.



Once the Beta scope and acceptance criteria have been agreed, I would therefore like a fixed quotation for bringing that defined Beta to completion.



Anything subsequently requested outside the frozen Beta scope can, of course, be separately scoped and quoted.



I am specifically looking for someone who is comfortable working in an AI-assisted/Lovable-first environment, where specialist developer time is concentrated on areas where professional technical expertise genuinely adds value.



If the Beta is successful and we work well together, there is a considerably larger Full Version/Post-Beta roadmap that I would expect to commission separately.

Please give me quote for this project

Kind regards



Luqmahn

Website management and design. Needed to do weekly updates as well as troubleshoot and test

About the Role

We're a growing luxury chocolate e-commerce brand looking for an AI Developer to help us build custom internal tools, quickly. We don't need a classic, slow-moving development cycle — we need someone who leverages AI-assisted development platforms (like Lovable.dev) to rapidly translate business needs into working dashboards and interfaces.

You'll work directly with our Shopify data and systems to build tools that give us fast, clear visibility into our business — sales trends, order operations, customer data, and more.

What You'll Do

Design and build custom dashboards and internal apps using Lovable.dev (or similar AI-powered/no-code development platforms)
Connect these tools to our Shopify store via APIs to pull and visualize order, product, and customer data
Work closely with our team to quickly understand business questions and turn them into functional data views (e.g., sales performance, inventory insights, customer segmentation)
Iterate fast — prioritize speed and clarity over heavy engineering overhead
Troubleshoot and refine tools based on ongoing feedback
Recommend the right AI/no-code tools for future data or workflow needs

What We're Looking For

Hands-on experience with Lovable.dev (required) or comparable AI app-building platforms (e.g., Cursor, Replit AI, Bolt.new, v0)
Experience integrating with Shopify's API (orders, products, customers, analytics)
Strong ability to translate business/data questions into simple, usable interfaces — you think like a product builder, not just a coder
Comfortable working independently, moving quickly, and iterating based on feedback
Bonus: experience with data visualization tools or dashboarding (e.g., charts, KPI trackers)
Bonus: past experience building tools for e-commerce or DTC brands

Why This Role

High ownership — you'll build tools our whole team relies on daily
Fast-paced, no bureaucracy — ideas go from concept to working dashboard in days, not months
Opportunity to shape our data/reporting infrastructure from the ground up

To apply: Please share examples of apps or dashboards you've built (ideally with Lovable.dev or similar tools), along with any relevant Shopify integration experience.

I’m looking for a Lovable Developer / AI App Developer to complete a very small $10 test task.

I want to create a simple AI web application using Lovable.dev. The first version only needs one page and a basic AI interaction.

Simple Task

You will:

* Build one clean page in Lovable.dev
* Create a responsive UI/UX
* Add a text input and button
* Connect it to a basic AI/API integration
* Fix any minor bugs

The app does not need complex features. I only need a functional AI MVP prototype.

Preferred Experience

* Lovable.dev
* Lovable AI
* AI App Development
* AI Web Applications
* Full-Stack Development
* MVP Development
* SaaS Development
* OpenAI
* AI Integrations
* API Integration
* React
* Next.js
* Supabase
* PostgreSQL
* JavaScript
* TypeScript
* Tailwind CSS
* Responsive Web Design
* Workflow Automation

Experience with Claude Code, Cursor, Replit, Figma to Lovable, and Figma to Code is a plus.

Budget

$10 fixed-price test

If the test goes well, I may have additional work involving AI applications, Lovable apps, SaaS platforms, MVPs, Supabase, React, API integrations, automation, and custom web applications.

To apply

Send one relevant Lovable / AI app / MVP / SaaS / full-stack example and briefly explain what you built.

Start your proposal with “AI LOVABLE”

Here is a more professional and technical version you can send to your development team:
PandaRide App – UI/UX Improvement Requirements
The core functionality of the PandaRide app is now working. The next priority should be improving the overall user experience (UX), user interface (UI), and ease of use.
The goal is to make the app simpler, faster, more intuitive, and visually professional, particularly for new users. Please focus on:
• Sign-Up & Onboarding: Simplify registration, reduce unnecessary steps, and provide clear instructions and progress indicators.
• Payment Setup: Make adding and managing payment methods straightforward, with clear confirmation when a payment method has been successfully added.
• Ride Request Process: Make requesting a ride simple and intuitive—from entering pickup/drop-off locations and selecting a vehicle to viewing the fare estimate and confirming the ride.
• Navigation & Screen Flow: Users should always understand what to do next. Minimize unnecessary screens, clicks, and repeated information.
• Appearance & UI: Improve the layout, spacing, buttons, icons, fonts, and overall visual consistency to give PandaRide a clean, modern, professional appearance.
• Error & Status Messages: Replace unclear loading screens or continuous spinning with meaningful messages such as “Processing,” “Payment Added Successfully,” “Searching for a Driver,” or “Documents Under Review.”
• Overall Objective: A first-time user should be able to create an account, add a payment method, and request a ride without needing instructions or assistance.
The app is functional; the focus now should be on refining and simplifying the customer experience so PandaRide is ready for public launch.