Full-Stack × AI Agents Engineer

We are scaling two highly ambitious lanes: an elite client-facing AI-specialist agent platform and a proprietary, end-to-end AI-native recruiting product. To accelerate, we need an engineer who can move past manual syntax-writing and orchestrate complex agentic systems that run, self-correct, and deliver flawless real-world utility.

You will own the complete technical lifecycle of our AI agent platforms, spanning a Python agent runtime, TypeScript/Node.js APIs, and Next.js interfaces.

• AI-Native builder: Directs agent workflows from a custom terminal or CLI-driven harness.
• Production LLM veteran: Shipped LLM systems to real users with robust evals and guardrails.
• Full-stack practitioner: Over 5 years of experience across TypeScript, Python, and modern web frameworks.
• Plain-language communicator: Explains technical trade-offs clearly and values direct, ego-free debates.

The company operates in a founder-led, fast-decision, and highly autonomous environment. Decisions are made without committee gates, using automated reviews instead, and the day-to-day rhythm is async-first and meeting-light to fiercely protect flow state. Communication is direct: team members are expected to disagree clearly, commit fully, and explain tradeoffs plainly. Quality is owned end to end with a bias to action ("try it, then discuss") and weekly demos directly to the founders where built work must perform. No sacred cows, including their own architecture, encouraging a "no sacred cows" philosophy where requirements are openly questioned.

• Pure flow state: Meetings are kept to an absolute minimum with an async-first workflow.
• Direct founder access: Run weekly demos straight to the founders and see your work ship instantly.
• Extreme autonomy: Automated code reviews and zero committee gates mean you own decisions end-to-end.
• Zero sacred cows: Freedom to challenge existing architectures and rewrite codebases for optimal performance.

TypeScript · Python · LLM system design · Evals & Guardrails development · Prompt Engineering · Tool-calling / MCP servers · CLI / Agent harnesses · Next.js · Node.js

This probably isn't for you if:

• You need consensus: We operate on a 'try it, then discuss' bias to action without committee approvals.
• You want to write every line of code manually: We expect you to leverage agents to write the bulk of your syntax.
• You struggle with direct feedback: We fiercely debate ideas, disagree openly, and commit fully to the chosen path.

If you are ready to build the future of agentic workflows alongside a highly focused team, we want to hear from you. Apply today with your GitHub or a link to tools you have built.