Python + LLM Engineer + Integration Role on Production AI Product

WHAT WE'RE BUILDING Production-grade Python AI product built on the Anthropic API (Claude Sonnet 5). The core classifier engine is live, deterministic, and running in a Docker container on a Hostinger VPS. We're moving toward WhatsApp Business integration, multi-tenant deployment, and expanded capabilities over the next 3-6 months. WHAT WE NEED Not a ticket-taker. We need a senior engineer who can look at where we are, look at where we're going, and add value we haven't thought to ask for. Concretely: someone who reviews architecture, catches design decisions that will cost us later, streamlines integration work, and brings expertise in production LLM systems we don't have in-house yet. When we hand you a piece of work, we want you to also tell us the three related things we should be thinking about but aren't. Founder continues to lead the current build phase. This role is advisory + delivery on emerging work, not replacement of core authoring. TECH STACK - Python 3.11+, uv package manager, pytest - Anthropic API (Claude Sonnet 5) - Docker on Hostinger VPS - Vercel serverless (Node and Python) - Upstash Redis (KV store) - Twilio WhatsApp Business API (upcoming) - Git (private GitHub repo, topic-branch + cherry-pick workflow) WHO WE'RE LOOKING FOR Required: - 5+ years Python. Production LLM integrations, ideally with the Anthropic API specifically. - Experience shipping WhatsApp / messaging channel integrations end-to-end. - Test-driven mindset. Every change lands under a test. - Git discipline. Small commits, descriptive messages, comfortable with cherry-pick workflow. Highly valued: - Experience working alongside AI coding tools (Claude Code, Cursor, etc.). We use them as primary authoring surfaces; you'd review and integrate their output, not just author from scratch. - Comfort with high-discipline codebases. We hold rigorous evidence contracts and byte-identical preservation. If "just ship it" is your default posture, we're not a fit. - POPIA / GDPR familiarity. HOW WE WORK Async-first. Slack for coordination, GitHub for review. Weekly sync call. Specifications precede code. TO APPLY Send: 1. Link to a production LLM integration you built (repo, case study, or client reference). 2. Your hourly rate. 3. One paragraph on a time you disagreed with an AI-suggested code change and how you handled it. Applications missing any of items 1-3 will not be reviewed. Contract: hourly, flexible hours to start, contract-to-hire potential based on fit.