Member of Technical Staff, AI Agents

As one of our first engineers, you'll shape the technical foundation of Socratix AI (YC S25). We build AI agents that investigate fraud and financial crime for banks, credit unions, and leading fintechs — agents that reason through real cases, gather evidence across systems, and produce decisions that stand up to regulators.

You'll design and ship production AI agents that fraud and compliance teams rely on every day, collaborate directly with founders on product strategy, and partner with both large enterprises and cutting-edge fintechs.

This is a high-impact role with massive scope for growth — you won't just be building features, you'll be defining how agentic AI transforms financial crime operations.

What You’ll Do

• Build agents end to end: LLM orchestration, tool use across a growing set of MCP integrations, retrieval, and multi-agent workflows that automate real investigations.

• Own the layer that makes agents trustworthy — eval harnesses, guardrails, audit trails, human-in-the-loop review. When your agent decides, a bank acts on it.

• Ship fast and close to users: code in prod your first week, direct customer contact, your work catching actual fraud.Design and deliver production-grade AI agents.

• Experiment relentlessly. New models, new techniques, new tooling — you'll have license to try everything and keep what wins.

• Turn frontier into product. You'll spend real time probing what the newest models can do, then ship the parts that survive contact with real fraud cases

• Shape roadmap and strategy with the founders. Small team, zero bureaucracy, maximum ownership.

What We’re Looking For

• A systems thinker. You reason about the whole machine — data flow, failure modes, latency, cost, what breaks at 10x — not just the function you're writing. Agents are distributed systems with a nondeterministic component in the middle; if that framing excites you, keep reading.

• You've shipped real things with LLMs — production systems, prototypes with users, obsessive side projects.

• Deep engineering fundamentals. Frameworks are details you pick up in a weekend (we're mostly Python on AWS). What matters is that you can design, ship, debug, and operate a production system end to end.

• 0 → 1 builder. Ambiguity and speed energize you instead of stressing you out.

• Roughly 2–4 years of engineering experience — or a portfolio that makes that line irrelevant.

• Prior experience in startups is a plus