Full-Stack Engineer (Python/FastAPI + React) with AI/LLM Experience — Ongoing Production Platform

# Full-Stack Engineer (AI / LLM) — Ongoing Production Platform **Remote · Work on a provided, sandboxed workstation · Long-term engagement** ## The role I run a licensed electrical contracting business, and I've built a custom AI-powered operations platform that runs the company — call intelligence, automated estimating, CRM, scheduling, and customer communications. It's **live in production with real customers and real data**, and actively growing. I'm looking for a **full-stack engineer with genuine AI/LLM experience** to work alongside me on both the frontend and the backend. This is an **in-flight project, not a greenfield build** — you'll pick up an existing, non-trivial codebase, coordinate closely with me, and ship improvements with me day to day. ## What you'll do - Work directly with me — plan, pair, and ship features together (frontend *and* backend). - Build and improve the dashboard UI and the APIs, data pipelines, and integrations behind it. - Work on AI features: retrieval (RAG), embeddings / vector search, prompt and agent design, and self-hosted LLM serving. - Fix production bugs, harden the system, and ship incrementally — with proper testing and clean rollbacks. ## Tech stack - **Backend:** Python (FastAPI), PostgreSQL, Docker - **Frontend:** React / Next.js, TypeScript - **AI:** self-hosted open LLMs (vLLM), embeddings + reranking, RAG, transcription, vision models - **Integrations:** QuickBooks Online, Twilio (SMS / voice), Gmail, Google Calendar, Cloudflare - **Workflow:** Git, Docker Compose, issue-tracked in Linear ## Required - Strong **full-stack** ability — comfortable across Python / FastAPI on the backend and React / Next.js + TypeScript on the frontend. - **Real AI / LLM experience** — RAG, embeddings / vector search, prompt engineering, and working with LLM APIs and/or self-hosted models. - Solid with **PostgreSQL, Docker, and Git**. - Able to **pick up and work inside an existing production codebase** — read it, understand it, extend it safely. - Able to **work on a provided, sandboxed virtual workstation** — you'll connect via remote desktop and VS Code Remote-SSH to our environment. All access is provisioned and isolated; you will **not** be working from your own machine. - A **clear communicator** who can coordinate closely with me, both async and live. ## Nice to have - QuickBooks Online API, Twilio, or Google (Gmail / Calendar) API experience. - Self-hosted LLM serving (vLLM), quantization, GPU inference. - Experience with agent / tool-calling architectures. - Linear or a similar ticket-driven workflow. ## How we work - **Ticket-driven and Git-based**, deployed via Docker. - I'm **hands-on and involved** — I set direction and coordinate scope with you, and you'll have real input on how things get built. - **Security is taken seriously** — scoped, sandboxed access and strict secrets discipline (e.g. credentials are never pasted into chats). ## Logistics - **Remote**, on a provided sandboxed workstation. - **[Hours — e.g. part-time / full-time]** - **[Rate / budget]** - **[Your timezone / overlap expectations]** - Ongoing, long-term engagement. ## To apply Please send: 1. A short note on your **full-stack + AI / LLM experience**, mapped to the stack above. 2. Links to **relevant work** (GitHub, shipped projects). 3. A line on your experience **working on remote/provided machines** and joining an existing codebase.