Part-Time Full-Stack AI Developer, AI Feature Maintenance (Construction SaaS)

About the project We're a small construction technology company building an internal platform that helps our team manage project estimates, site documentation, client communication, and project workflows. One of our core features is an AI assistant that analyzes construction documents, summarizes project information, and helps generate estimates and reports. The feature is already live and being used internally. We're looking for a hands-on Full-Stack AI Developer to maintain it, resolve a few known issues, and deliver a couple of focused improvements—not rebuild the platform or create a new product. What you'll actually be doing * Maintain and debug the existing AI document analysis pipeline (prompt logic, API calls, error handling) * Improve the accuracy and consistency of AI-generated project summaries and estimate recommendations * Add one small new capability: automatically flag missing project details, potential cost risks, or inconsistencies before reports are finalized * Light full-stack work to integrate the feature into our existing backend and frontend (no major architecture or infrastructure changes) * Use AI-assisted coding tools (Cursor, Claude Code, or similar) to work efficiently while keeping changes clean and well-tested Tech stack Python/FastAPI backend, React frontend, PostgreSQL, OpenAI/Anthropic APIs. We're more interested in your ability to understand and improve an existing codebase than matching every framework exactly. Who we're looking for * Proven experience shipping LLM-powered features in production applications (not just demos or side projects) * Comfortable debugging and improving existing AI workflows * Able to work independently and communicate asynchronously * Regular user of AI coding tools (Cursor, Claude Code, Copilot, etc.) to accelerate development without sacrificing code quality * Experience with construction, engineering, estimating, or document-processing software is a plus, but not required Deliverables by the end of the month * Existing AI feature bugs resolved * Risk/inconsistency detection feature implemented and tested * Improved reliability of AI-generated summaries and estimate suggestions * Short handoff document explaining changes, known limitations, and recommendations To apply Please share 1–2 examples of production AI/LLM features you've built and tell us your role in shipping them. We're interested in real-world applications rather than prototypes.