Member of Technical Staff

• OSINT experience: deep online investigations, identity resolution, username tracing, entity mapping. This one counts for a lot.
• Something you've built on LLMs or agents, like a research tool, a RAG app, a scraper, or an agent loop.
• Distributed systems, queues, workflow engines, or high-throughput pipelines.
• React or Next.js, especially data-heavy interfaces.
• Production experience with Postgres, Redis, search, and observability.
• A nose for answers that look right and aren't.

We don't care much about pedigree. If you took a non-traditional path or haven't used our stack, apply anyway; we'll teach you the tools. How you think about problems is the one thing we won't bend on.

• Full-time, in person in San Francisco.
• Level: new grad through senior.

Send us something you built and be ready to go deep on it: the design you chose, the alternatives you threw out, and what breaks first at 100x the load. We're a small team, so every hire changes what the company can do.

• Ownership of features that go to production and paying customers, with your name on them.
• Close review of your designs and code from engineers who build research agents, evals, and large-scale systems every day. Expect real pushback; it's how people get good here.
• Room to scope your own work, make the call, and watch it land.
• A generous budget for frontier models and dev tools. Use the best ones.
• Competitive salary and meaningful equity.
• Full Health, dental, and vision insurance
• Lunch in the office and team offsites.

Sixtyfour builds research infrastructure for decisions that can’t afford to be wrong. Our AI agents investigate people and companies across public records, web pages, archives, and other sources, returning structured findings with supporting evidence and a graph of the connections between them.

We help teams investigate fraud, research the people behind accounts, and understand who they’re doing business with. Start with an email, a username, or a company domain. Sixtyfour follows the evidence, connects records across sources, and gives teams findings they can inspect and use in their workflows.

The hard problems are in making that research reliable at scale. Do two profiles belong to the same person? What happens when sources disagree? Does a shared identifier reveal a meaningful connection? How do you investigate millions of records while keeping latency and cost under control? These are the problems we’re building for.

We’re backed by Y Combinator (X25)

How we work

We design before we build. On most projects, the bulk of the time goes into understanding the problem, sketching the system, and arguing through failure modes and tradeoffs at a whiteboard. Once the design holds up, the code goes quickly, and customers usually see it within weeks.

We keep a high bar, and everything we ship has to hold up at production scale. If the whiteboard is your favorite part of engineering, you'll like it here.