Freelance python jobs for world-class Developers

I’m building AIVI QuickScan (starting with dentists). The core pipeline already works: Python fills a Jinja HTML template in memory, then Playwright renders a PDF. I need an expert to make the PDF output feel premium AND remain stable across varying lengths.

NOTE: This is NOT a Figma-only design role. You must implement print-safe CSS and prove stability on long PDF output.

========================
MILESTONE A (FIXED PRICE)

You’ll work in a private GitHub repo and run a local command that generates 3 PDFs (short, medium, long) from sample data.

DELIVERABLES

Shared print-safe design system

Create: styles/fixpack_base.css (typography scale, spacing system, reusable components)

Components should cover: cards, chips or badges, checklist rows, callouts, section headers

Include print-safe rules for pagination and keep-together blocks

Upgrade QuickScan report styling

Update the report template and CSS to use the shared design system (minimal template edits preferred)

Improve typography, hierarchy, spacing rhythm, and consistency

Must look good and remain stable specifically on quickscan_long.pdf

Fix Pack FAQ template (SA5)

Create: templates/fixpacks/fixpack_faq_cards.html using the same design system

Must match the report style (same components and typography)

Stability proof (required)

Provide stable renders on short, medium, long PDFs:

no orphan headings

no cut-off content

avoid split cards or checklist blocks (or split only with a controlled pattern)

2 revision rounds included

NOT IN SCOPE

No scoring logic changes

No building a SaaS or web app

No Stripe, admin, or keys

No monitoring or subscriptions

REQUIREMENTS (MUST HAVE)

Strong HTML and CSS plus typography

Proven experience with HTML and CSS to PDF rendering (Playwright or Puppeteer) plus print CSS

GitHub workflow (branch and PRs)

Can run local render command and iterate quickly

TO APPLY

Share 2 to 3 examples of HTML and CSS to PDF work (before and after strongly preferred)

In 3 to 5 bullets: how you handle page breaks, orphan headings, and keeping cards together

Confirm you can run the repo locally and submit changes via PR

If this goes well, there will be follow-on milestones (more fix packs, report extensions).

We are seeking an experienced Python developer to assist in deploying our emergent lab application on the web (App is already done). The ideal candidate will have a strong background in Python programming and web deployment strategies. You will work closely with our team to ensure the application is seamlessly launched and operational. Familiarity with web frameworks and cloud services is a plus. If you are passionate about technology and ready to tackle new challenges, we want to hear from you!

We’re looking for a developer to build a manual task orchestration system for our Vinted operations. Due to API restrictions, our team must upload and manage listings manually, so we need a web-based system that assigns tasks, tracks progress, and minimizes human error.

The system will include a queue that distributes tasks (uploads, reuploads, repricing, deletions, etc.) to around 10 concurrent staff members, a simple task interface with proof-of-completion uploads, priority handling for critical actions, and a dashboard for performance tracking. It will also integrate with email parsing for sales, Shopify, and our internal systems.

The main goal is to handle 1,000+ daily uploads and reuploads while keeping error rates under 1%. We already have a detailed specification and workflow; we’re looking for someone experienced in building internal tools, task queues, and admin dashboards.

Please include examples of similar systems you’ve built, your recommended tech stack, and an estimated timeline for an MVP.

Full spec attached

We are seeking an individual with extensive experience in programming AI bots. Your expertise in this specialized area would be greatly valued. We are eager to connect with someone who possesses a strong understanding of AI development. Please let us know if this opportunity aligns with your professional background. We look forward to hearing from you.

We are hiring a senior engineer to own the design, implementation, and reliability of a production AI agent that communicates over messaging channels (including WhatsApp and iMessage), supports multiple languages, and triggers real backend actions through tools.

This role is focused on deterministic agent behavior, strong tenant isolation, memory design, and multilingual correctness.

Responsibilities

-Build a multi-channel ingestion and normalization layer (WhatsApp + iMessage) into a unified internal message schema
-Implement robust identity resolution across channels (phone, email, platform handles) with tenant-scoped deduplication
-Design and implement a deterministic tool-calling architecture with structured outputs, validation, idempotency, and permissions
-Implement short-term conversation state and long-term memory with strict tenant isolation
-Implement vector-based retrieval for broker/tenant knowledge, templates, summaries, and document RAG with hard scoping and RLS
-Implement multilingual support using a pivot-language strategy, including translation, locale formatting, and RTL handling
-Reduce hallucinations and repeated responses through architecture, not prompts
-Add observability, tracing, and automated evals for key failure modes (wrong-tenant retrieval, incorrect tool calls, repetition, hallucinated facts)

Technical environment
-Node.js and TypeScript
-Postgres (Supabase) and vector search (pgvector or equivalent)
-Redis for short-term state and rate limiting
-LLM APIs and tool-calling loops
-CI-based regression testing and evaluation

Required experience
-Proven track record shipping LLM agents in production that call tools and take real actions
-Multi-tenant architecture and data isolation experience
-RAG/vector retrieval implementation with strict filtering
-Experience building messaging integrations or event-driven systems
-Strong testing and observability discipline

We are looking to encrypt the eMMC and Micro SD cards on our Raspberry Pi projects. We would prefer to use an external cryptographic co-processing chip, such as the ATECC608B, in order to encrypt the device storage without needing a password when the device reboots.

With an external co-processing chip, we hope to reduce any additional overhead requirements on the Raspberry Pi devices.

The Raspberry Pi devices will either be running Raspberry Pi OS or Debain/Ubuntu.

The final deliverable will be a script that we can run on a Pi 4/CM4, Pi 5/CM5, Pi Zero W and Zero 2 W to automate this process.

We are seeking an expert in Python and Playwright (or similar frameworks) to assist with a two-phase data integration project for an enterprise Document Management System (DMS).

Phase 1: Security & Resilience Testing

Perform environment resilience testing and implement AI-driven session management to handle enterprise-level security protocols.

Provide professional recommendations on fixing security hurdles related to firewall stability and environment detection.

Phase 2: Data Synchronization & Lead Generation

Build a synchronization tool using AI/LLM-assisted data extraction to ensure high accuracy when moving lead data to our Bubble.io backend.

Ensure seamless data transfer from the DMS interface into our Bubble.io backend.

Technical Requirements:

Expert-level proficiency in Playwright, Selenium, or Puppeteer.

Experience with complex session persistence and error handling (403/401).

Strong understanding of API integrations (Webhooks / Bubble.io).

Note: A full 16-page SOW is ready. For privacy and security reasons, this will be shared directly with shortlisted candidates in the message room.

• Which specific AI models or libraries do you recommend for handling adaptive environment detection in high-security DMS systems?
• For a high-security DMS that uses dynamic session tokens and environment-based fingerprinting, would you recommend a 'Stateful' or 'Stateless' browser context in Playwright, and why?
• How do you handle 'Canvas' or 'WebGL' fingerprinting when automating inside a sensitive enterprise portal without triggering a 403 error?
• When piping data to a Bubble.io backend, how do you handle rate-limiting or partial-failure states to ensure no data is lost during the sync?
• How would you integrate an LLM (like GPT-4o or Claude) into a Playwright workflow to handle 'self-healing' selectors if the DMS interface undergoes a minor layout change?

Overview

We are a flooring retail business based in Mackay, Queensland (Australia). Each month we export data from our business operating system (RBOS/FLEXI style reports) into Excel. We want an AI-enabled reporting agent that takes the monthly exports as input and automatically generates a consistent pack of sales performance reports in a well-formatted Excel workbook.

The agent must merge data across multiple exports using a shared Quote/Order ID, apply filters (including customer type = Residential where required), calculate KPIs, run quality checks, and output a polished report pack that is easy to read.

We will upload the monthly exports manually and run the agent monthly.

Reports Required (Must Generate)

The agent must generate the following reports each month:

Report 1: Monthly Sales Performance (Residential only)

Per salesperson:

Orders count

Total revenue

Total GP $

GP % (GP $ / Revenue)

Average job value

Average GP per job

Plus a “Residential Data” tab with row-level detail.

Report 2: Quote to Order Conversion (Residential only unless agreed otherwise)

Per salesperson:

Quotes issued (count, value)

Orders won (count, value)

Conversion % by count

Conversion % by value
Rules:

Status = Quote means not won

Status = Lost means lost

Status = Order means won

Must be calculated from exports provided

Report 3: Open Quote Pipeline (Residential only unless agreed otherwise)

Per salesperson:

Open quotes value and count

Age buckets: 0–14, 15–30, 31–60, 60+ days

Optional predicted revenue based on trailing conversion %

Report 7: Average Time to Convert

Per salesperson:

Median days from quote to order (exclude same-day conversions)

% converted within 7 days

% converted within 14–30 days

Report 8: Quarterly Growth Report

Per salesperson:

Revenue growth vs previous quarter

GP growth vs previous quarter

Average job value trend vs previous quarter

Report 11: Job Size Distribution

Per salesperson:

Job counts and revenue split by bands:

Under $5k

$5k–$15k

$15k–$30k

$30k+

Report 13: Annual Revenue & GP Ranking

Rank salespeople by:

Total revenue

Total GP $

GP %

Orders count
Include comparisons vs previous year if data provided.

Inputs (What We Provide Monthly)

One or more Excel exports from our RBOS/business operating system (FLEXI-style reports).

Typical exports include:

Sales by salesperson with profit/GP

Orders by customer type

Quotes/orders status export

Exports will contain a shared Quote/Order ID used to link rows across sheets.

The solution must be robust to minor column name changes and different sheet names, as exports can vary.

Output Requirements

A single Excel workbook generated per month:

Clearly named tabs for each report

Clean formatting: bold headers, light header fill, borders, frozen panes

Currency formatting ($#,##0.00), percentage formatting (0.00%)

Auto column widths

Report pack must be consistent month-to-month

Include “Data” tabs where appropriate (traceability)

Include a summary dashboard tab with key highlights (optional, but preferred)

Required Features
Data Handling

Automatically detect the correct join key (Quote/Order ID)

Merge datasets safely (no duplicate inflation)

Filter Residential-only where specified

Handle blanks, cancellations, and invalid values

Ensure numeric parsing for revenue and GP fields

Quality Control Checks (Must Include)

Check residential filter has no leakage

Check totals reconcile: summary totals must match filtered raw totals

Flag duplicates by Quote/Order ID

Flag missing required columns with clear error messages

Usability

We want a simple monthly process:

Drop new exports into a folder

Run the agent

Get the Excel report pack output

Preferred interface:

Windows-friendly local app (simple GUI), OR

Command line script with clear instructions, OR

Google Colab style notebook (only if it’s dead simple)

The solution should be maintainable without developer involvement every month

Tech Preferences

Open to recommendations, but likely options:

Python (pandas + openpyxl) for ETL + Excel formatting

Optional: an “AI layer” for column mapping, anomaly detection, and narrative insights

If using an LLM, it must not break determinism for core calculations (AI can assist, not replace math)

Deliverables

Working AI reporting agent that generates all required reports

Clean, commented codebase

Installation/setup instructions

A short user guide: “How to run it monthly”

Example output workbook using our sample exports

A configuration file for mapping columns and report rules (preferred)

Acceptance Criteria

Reports 1,2,3,7,8,11,13 generated correctly from sample data

Residential-only filtering correct where required

Output Excel file is polished and easy to read

Reconciliation checks pass

Can run monthly with minimal steps and no debugging

To Apply, Please Include

A brief outline of your approach

Examples of similar reporting automation work

Recommended stack and why

Estimated timeframe and fixed price or milestone breakdown

Suggested Milestones (So You Can Manage Risk)

Milestone 1: Data ingestion + robust merge by Quote/Order ID + Residential filter + Report 1 output

Milestone 2: Reports 2 and 3 + validation checks

Milestone 3: Reports 7, 11 + formatting polish

Milestone 4: Reports 8 and 13 (quarterly + annual logic)

Milestone 5: Packaging (GUI or run-script), documentation, final testing

Job Title
High-Frequency Trading Bot Developer – Rithmic Time & Sales (Order Flow) – MT5 Execution

---

Job Description

I am looking for a serious and experienced algorithmic trading developer to build a high-frequency scalping bot based primarily on Time & Sales (executed trades) using Rithmic data, with execution on MetaTrader 5 (IC Markets – US500).

This is not an indicator-based bot, not a grid, not martingale, and not a demo/UI-only project.
I need a real working trading system, inspired by professional order-flow / HFT concepts, adapted to a retail execution environment.

---

## Project Objective

Build a second trading bot focused on Time & Sales order flow, capable of generating approximately 500–800 trades per day (600+ target) during liquid market hours.

The bot must:
- Analyze executed trades only (Time & Sales)
- Detect aggressive trade bursts
- Trade micro price movements (1–2 ticks)
- Exit very fast (seconds)
- Respect IC Markets rules

---

## Data Source (Mandatory)

- Rithmic Time & Sales (Last Trade feed)
- Each trade includes:
- timestamp (milliseconds)
- price
- volume
- aggressor side (BUY at ask / SELL at bid)

DOM Level 2 is optional, not required.

---

## Strategy Logic (Must Be Implemented)

### 1) Market Context (Average Flow)

Continuously compute a rolling average over the last 3–5 seconds:
- trades per second
- BUY vs SELL balance
- average delta (BUY – SELL)

This average is used as context, not as a trigger.

---

### 2) Burst Detection (Main Trigger)

Detect bursts inside a short rolling window (250–500 ms).

#### SELL Signal (example):
Enter SELL if ALL conditions are met:
- ≥ 20 SELL trades within 500 ms (configurable)
- SELL trades ≥ 2× BUY trades
- SELL volume ≥ 70% of total volume
- Trade speed significantly higher than average
- Price already moves –1 tick

BUY signal is the exact inverse.

---

### 3) Decision Rule

Trade only if:
- Burst direction matches the average market bias
- Spread is below threshold
- No news / abnormal volatility
- Cooldown respected

This avoids trading noise and fake pressure.

---

## Trade Management

Each trade:
- Entry: market or aggressive limit
- Take Profit: +1 to +2 ticks
- Stop Loss: –1 to –2 ticks
- Timeout exit: close trade if no movement after 2–5 seconds

No holding, no averaging, no martingale.

---

## Frequency Target

- 500–800 trades/day on average
- Higher during US session
- Small position size
- High-frequency, low exposure

---

## Execution & Architecture

- Execution on MetaTrader 5 (IC Markets)
- Strategy engine can be:
- C++
- Python
- or hybrid
- Communication with MT5 via:
- DLL
- Socket
- ZeroMQ
- or equivalent low-latency bridge
- Must be VPS-friendly (New York / low latency)

---

## Required Skills

- Algorithmic trading experience
- Time & Sales / Order Flow strategies
- Rithmic API (or similar institutional feed)
- Low-latency systems
- MT5 automation (MQL5 or bridge)
- Understanding of spread, slippage, execution risk

---

## What You Must Explain in Your Proposal

1. Have you built Time & Sales–based bots before?
2. Have you worked with Rithmic or similar data feeds?
3. How would you detect bursts efficiently?
4. What language / architecture do you recommend and why?
5. Estimated price ($) and timeline for:
- MVP
- Stable production version

If this type of strategy is not your specialty, that is totally fine — just be honest.

---

## Important Notes

- No martingale
- No grid
- No indicators
- No marketing bot

This is a serious technical HFT-inspired project, adapted for retail execution.

Looking forward to working with a professional developer.

Best regards,
Saïd

Project Scope and Objectives:

the target website URL: https://app.airdna.co

login requirements: create your own free account

exact data fields to extract :

at the market level (airDNA / find a market)
* market score
* annual revenue
- at the market level (airDNA / find a market / listings)
* total active listings
* total active listings; growth rate

at the market level (airDNA / find a market / submarket)
- market score

volume/frequency: one-time
output format: CSV
sample data: https://docs.google.com/spreadsheets/d/12vgIFmGFOgSUlm9e750ho3aV4yNXgoJub2Nv9f_LQOQ/edit?usp=sharing

Budget, Timeline, and Payment Structure
- Milestone 1: send test data before completing project
- Budget: please provide your best offer
- Timeline: please let me know when this could be completed by. people who can finish fastest will be prioritized

Technical lead / founding engineer for an applied AI fintech venture.
The team is targeting someone who combines strong product engineering with real data systems experience.

Profile:
- Senior full-stack engineer (5–7+ years) from a strong fintech or infrastructure company (Stripe / Plaid / Brex–type environment or similar caliber)
- Strong in React, TypeScript, and Python
- Comfortable building end-to-end systems
- Experience working with financial data, security, or compliance-heavy systems
- Experience designing data pipelines and handling non-trivial data volumes
- Comfortable building scoring / analysis systems (deterministic + model-assisted logic)
- Strong understanding of AI/LLM capabilities, trade-offs, and evaluation frameworks (not just API usage)
- Thinks in terms of system reliability, observability, and cost control
- Pragmatic builder — ships fast but with discipline
- High ownership mentality; proactive and product-aware
- Ideally interested in fintech / wealth management / B2B SaaS
- Master’s/PhD in CS (or similar) is a plus, but execution and commercial instincts matter more. Open to academic backgrounds if there’s applied, production experience.
- Ownership. Speed. Innovation. User Centered. Analytical and business oriented.

Looking for someone who wants meaningful ownership and long-term upside. Initially Part-time and milestone driven.

#collaboration #part-time

Rate depends on seniority & deliverables.

We’re hiring a Senior AI Demand Forecasting Lead to design and own a single, intelligent demand forecast across a complex omnichannel commerce ecosystem.

This is not a reporting role.
This is not a traditional planning role.

This person will build the forecasting brain that powers buying, production, replenishment, and inventory risk decisions across:

DTC eCommerce (Shopify)

9 owned retail stores

Multiple marketplaces flowing through Shopify

Amazon

Drop-ship marketplace(s)

Franchise locations operating on separate POS systems

The challenge: fragmented data, different fulfillment models, different inventory ownership structures — all needing to roll into one SKU-level forecast that reflects true demand, not just booked sales.

What You’ll Do
🔹 Build a Unified Forecasting System
Design an end-to-end demand forecasting architecture that consolidates all revenue channels into a single source of truth

Normalize and reconcile sales data across:

Owned inventory

Drop-ship demand

Franchise demand

Marketplaces with delayed or partial reporting

🔹 Apply AI & Advanced Forecasting Methods
Build SKU-level forecasts using:

Time-series models (Prophet, ARIMA/SARIMAX)

Machine learning models (XGBoost, LightGBM, LSTM where appropriate)

Incorporate:

Seasonality

Launch timing / time-on-shelf normalization

Promotion and discount lift

Channel-specific behavior

New SKU analog logic using attributes

🔹 Channel-Aware Demand Logic
Separate true demand from inventory-constrained sales

Avoid double-counting demand across channels

Translate external demand signals (Amazon, drop-ship, franchise) into internal production and buy recommendations

🔹 Data Engineering & Integration
Build automated pipelines pulling from:

Shopify APIs

Amazon Seller/Vendor data

POS systems (Lightspeed, external Shopify)

Existing Excel-based planning tools (as a bridge, not a crutch)

Reduce manual Excel dependency while respecting current workflows

🔹 Business Partnership
Work closely with Product, Merchandising

Explain forecasts in business language, not just models

Build trust and adoption across teams

What We’re Looking For
Must-Have

7–10+ years in demand forecasting, data science, or advanced analytics

Hands-on experience building forecasting models (not just consuming tools)

Strong Python skills (pandas, numpy, sklearn, statsmodels)

Experience working with messy, real-world commerce data

Omnichannel experience (DTC + retail + marketplaces)

Strongly Preferred

Shopify and Amazon data experience

Retail inventory mechanics (on-hand, on-order, lead times, base stock)

Experience modernizing or replacing Excel-based forecasting systems

Consumer goods or high-SKU environment experience

What This Role Is Not
❌ Not a reporting analyst

❌ Not a BI/dashboard-only role

❌ Not a traditional retail planner

❌ Not a “load data into a tool” job

This role builds the system.

Why This Role Is Different
High ownership and visibility

Real-world forecasting problems (not academic exercises)

Complex, interesting data

Direct impact on inventory risk, margin, and growth

Opportunity to architect something foundational

***Read this entire document before applying. Your proposal must reference specific details from this posting. Generic proposals will be rejected immediately.***


The Situation
We are a profitable precision nutrition meal delivery service. Entering an aggressive growth phase in 2026.

Our entire technology infrastructure has been managed by a single engineering contractor for three years:

One person controls all hosting, database access, credentials, and deployment pipelines
There is no documentation. No architecture diagrams. No runbooks
We do not know our hosting provider, configuration, or redundancy setup
We do not have a complete inventory of services, credentials, or integrations
No monitoring. No alerting. No disaster recovery plan
10+ years of customer, order, and financial data in a PostgreSQL database that no one on leadership can access directly
The contractor consistently exceeds budget caps without flagging overruns and slow-rolls access requests

This is an existential risk. You are the person who maps, secures, and hardens the infrastructure so it is no longer a black box controlled by one person.

You will work alongside a Senior Full-Stack Engineer (being hired in parallel). They own the application layer. You own the infrastructure layer. Together you eliminate the single point of failure and give leadership full visibility into the systems that run the business.





The Stack
Frontend: Custom-built application
Backend: Custom application layer, Rails and/or Node.js
Database: PostgreSQL, 10+ years of production data
Hosting: Cloud-hosted. Provider and configuration unknown
Migration: Moving to Shopify. Infrastructure must support the migration and post-migration architecture
Integrations: FedEx shipping, payment processing, SMS/email marketing, various operational tools. Full inventory unknown





Your First 30 Days: Concrete Deliverables

Deadline
Deliverable
Day 5
Complete infrastructure map: every server, database instance, DNS record, SSL certificate, CDN, load balancer, cron job, and third-party service. Where it runs, who pays for it, and who has the credentials.
Day 10
Credential and access audit: every account, API key, SSH key, and password. Where they are stored, who has them, which are shared, which have MFA, and which are single-person dependencies. Severity-rated risk matrix.
Day 15
Backup assessment: what is currently backed up, how often, where backups are stored, and whether anyone has ever tested a restore. If no verified backups exist, flag it as critical and propose a 48-hour fix plan.
Day 20
Monitoring and alerting MVP: basic uptime monitoring, database health checks, error rate tracking, and disk/CPU/memory alerts.
Day 30
Disaster recovery playbook: documented, tested with a restore drill, written so a non-DevOps engineer could execute it in an emergency. Includes RTO/RPO targets and escalation procedures.






After 30 Days
Credential management. Vault, AWS Secrets Manager, or equivalent. No more credentials in plaintext or shared accounts.
CI/CD pipeline. Staging environments, automated deployment, and rollback capability.
Infrastructure as code. Reproducible, version-controlled infrastructure. Nothing lives only in someone's head.
Migration support. Ensure hosting, database, and services are ready for the Shopify migration and post-migration architecture.
Cost optimization. Review and right-size all infrastructure spending. Present findings with recommendations.
Scale planning. Capacity modeling for 2x and 5x current volume. What breaks first and what to invest in before it does.





Requirements
Non-Negotiable
4+ years in DevOps, SRE, or infrastructure engineering. You have owned the reliability of production systems.
Deep experience with AWS, GCP, or Azure: provisioning, networking, IAM, security, cost management.
PostgreSQL administration: backup, replication, performance tuning, access control.
You have audited or taken over infrastructure from someone else with minimal documentation.
You write runbooks, architecture diagrams, and incident response procedures as standard practice.
You can explain infrastructure risks to non-technical stakeholders in plain language.
Upwork auto-tracked time. No exceptions.

Strong Preference
Docker, Kubernetes, or container orchestration experience.
Terraform, CloudFormation, Pulumi, or other IaC tools.
Experience supporting e-commerce platform migrations (Shopify or similar).
Data pipeline or ETL experience (BigQuery, Metabase, Fivetran).
Incident response and on-call experience.
DTC, subscription, or food tech background.





Automatic Disqualifiers
Do not apply if any of these are true:

You have never audited or inherited someone else's infrastructure
You do not document your work as you go
You are uncomfortable sharing access or credentials with team members
You cannot explain infrastructure concepts without jargon
You are not willing to use Upwork auto-tracked time





About the Company
$26M revenue, profitable, zero outside debt
16,000+ meals produced weekly across commercial kitchens
Nationwide DTC delivery in reusable glass jars
Precision nutrition with measurable health outcomes
2026: new channels, product lines, significant operational expansion
Lean team, high standards, direct communication, zero bureaucracy





How to Apply
We know most applicants will use AI to write proposals. That is fine. What is not fine is proposals with zero evidence of actual experience. Your screening question answers can only be good if they come from someone who has actually done this work.

Answer every screening question. Reference specific details from this document. Proposals that could apply to any DevOps role at any company will not be read.

What Happens After You Apply
30-minute video call. You walk through your screening answers live. If the details get vaguer than what you wrote, the interview ends.
Paid technical exercise (2-4 hours, at your rate). Access to a test environment. Produce a mini infrastructure audit: map what you find, assess risk, document findings for a non-technical audience.
Offer or pass. If the exercise confirms your application, we move to contract. If not, we pay you for the exercise and part ways.

No whiteboard questions. No trivia. We pay you to do a small version of the job and evaluate the output.