Profile description: You set the technical direction and quality bar for an organization where autonomous AI agents generate most of the code. You decide what gets built and how it's shaped, you judge the AI's output as the quality gate the system answers to, and you push the agentic system itself to the next level. One part principal software architect, one part applied-AI systems engineer. Developer profile: Own architecture and technical taste across the products being built — the system design, the hard tradeoffs, the "is this the right shape" call. Judge the AI's output: review agent-generated work for correctness, design, and quality, and define the standards the engine is held to. Your judgment is the bar. Advance the engine itself: improve how agents are orchestrated, prompted, and combined — multi-agent design, model selection, reliability of the AI pipeline. Take on novel, greenfield system design for products going to the wild, including the hard parts AI can't yet own. Shape engineering culture — how humans and AI agents build together, raising the bar for everyone. Technical requirements: 10+ years of software engineering, including senior/principal/staff-level architecture ownership Deep full-stack and systems design experience, with strong, well-held opinions on what good looks like Strong applied-AI experience with LLMs and agentic/multi-agent systems — orchestration, prompting, tool-use, reliability. You've built or seriously operated AI systems, not just used them Exceptional judgment on code quality and design — you can look at output and instantly see what's wrong, fragile, or elegant Comfortable with high autonomy and ambiguity; calibrated divergence — you see differently, push back, and converge to the right answer Bonus: production experience with multi-agent/autonomous systems, security and data-boundary instincts, SOC 2 or regulated-environment awareness, and a track record of taking AI capability to the next level in a real product Required tests: To be confirmed with the hiring lead — expect a system design/architecture round and a working session evaluating and critiquing AI-agent-generated code. ATTENTION UPON DROPPING YOUR APPLICATION: Please, immediately send an email with subject 'Principal Engineer/Architect – Your Name' sharing concrete examples of software architecture and applied-AI/agentic-systems work you've done in production. We're specifically interested in: Production systems you've architected or owned end-to-end — the scale, the stack, and the hard tradeoffs you made LLM or agentic/multi-agent systems you've built or operated — how they were orchestrated, prompted, and made reliable in production, not in a demo Times you reviewed or gated AI-generated code — what standard you held it to, and what you caught that a less rigorous reviewer would have missed Multi-agent orchestration, model selection, or reliability work on an AI pipeline — the problem it solved and how you measured success Greenfield architecture decisions for a product that actually shipped to real users — the "is this the right shape" calls you made and why Any security or data-boundary work in a regulated or SOC 2 environment, if relevant Please point us to production work, not tutorials, side experiments, or course projects. A short paragraph per example is enough: what you built, your specific role, the scale or context it operated in, and one or two concrete decisions you made. Links to repos, demos, or write-ups are welcome where you can share them.