Founding Platform Engineer – Agentic Systems

We're a small, high-velocity team building real agentic infrastructure. Not hype — actual swarms that ship, fail, learn, and compound. Your first mission: own the internal backbone for our team (tenancy, vaults, agent definitions, ops visibility). This is deep platform work with serious AI orchestration depth. We want founding-level engineers who’s already lived the 100x–1000x agentic world and is excited to take risks, own big surfaces, and move fast. You treat dozens of agents as your team and know how to make them reliable at production scale. The Mindset We’re After You’ve shipped real agentic systems in production (scars included). You obsess over why things broke and how to make the next version 10x more robust. You think in systems, not frameworks. You can go deep on LLM failure modes, evals, memory, orchestration trade-offs, and compounding loops. Builder + operator energy: you get bored in big companies, push back when something feels off, and love owning a domain end-to-end. Risk-tolerant founder mentality: willing to experiment, iterate fast, and bet on bold agentic patterns. Strong synergy with a small team that moves at founder speed. Technical Depth (Bullet-Point Reality Check) Real agentic production experience — multi-agent swarms (not demos, not basic RAG, not “played with LangChain”). Tell us the worst bug you shipped and how you caught it. LLM systems thinker — prompt regressions, CAG vs RAG, tool-use failures, deterministic vs stochastic orchestration, evals you’ve actually designed. Strong Python backend — FastAPI, SQLAlchemy, async, and especially Postgres (RLS + audit tables). You care about data integrity at the DB layer. Frontend capable — can ship clean Next.js UIs that just work without a designer. Infra literate — ECS, RDS, Cognito, CloudFront, Docker, GitHub Actions. You debug weird production behaviors. Orchestration & compounding — experience evolving agent swarms that get smarter over time (memory, review loops, self-improvement). Hard Filters (Please don’t apply if this is you) ML researcher / PhD focused on fine-tuning models “AI-curious” backend dev with no shipped agentic work Prompt engineer without real software engineering depth Frontend-heavy full-stack who mostly just calls APIs Whiteboard architect who doesn’t ship Anyone who leads with framework names instead of systems reasoning How We Screen (Fast & Honest) This role is rare and high-impact. If you’ve built and battled real agent swarms, love taking ownership, and want to help shape the platform that powers exponential team leverage — apply with: Resume or LinkedIn 1–2 examples of agentic work (repo, outcomes, or war story) Let’s build something that actually feels like the future.