// HACKER NEWS — CYBERSECURITY
Sixtyfour (YC P25) Is Hiring
The Opportunity
Sixtyfour turns a single name, email, or domain into a full, verified picture of a person or company — by sending AI agents out to research the open web the way a sharp analyst would, then checking and scoring what they find. You'll build real parts of that: the agents that reason and gather evidence, the systems that run them at scale, and the product people use to see the results.
How We Work
We spend most of our time — call it 80% — understanding the problem deeply, planning, and designing the system before a line gets written. Getting the design right is the hard part and the best part. We hold a high bar, we stay on the edge of what's possible, and everything we ship has to hold up at scale. If you love the part of engineering that happens on the whiteboard — arguing the right design, the failure modes, the tradeoffs — you'll fit here.
What You'll Do
- Build AI agents for OSINT and deep web research — design agents that investigate people and companies across the open web, public records, social platforms, and other sources, then cross-reference and structure what they find.
- Own the thinking, not just the code — dig into the problem, weigh the designs, and write the plan before you build, because that's where the real leverage is.
- Design systems that hold up at scale — reason through data volume, concurrency, latency, and cost up front, so what you build survives real load.
- Ship features end to end — design, build, test, deploy — so customers get something new in weeks, not quarters.
- Build and sharpen the AI agents that research people and companies, so enrichment returns more accurate, better-sourced answers.
- Add new data sources and tools to the enrichment engine, so agents can reach information they couldn't before.
- Write evals and tests that prove whether a model or agent change actually made results better, so the team improves on evidence instead of hope.
- Make long-running jobs fast and reliable — batching, caching, retries, orchestration — so millions of records enrich without falling over.
What We're Looking For
Must-have — this is a high bar, and we mean it:
- Strong engineering fundamentals. You understand how real systems work underneath — concurrency, APIs, databases, how the web fits together — and why they're built that way. Syntax is the easy part; you get the concepts beneath it.
- System-design and architecture instinct. Hand you a fuzzy problem and you can break it into pieces, find the failure modes, weigh the tradeoffs, and design something that holds. You think before you build.
- You think at scale by default. You reason about data volume, concurrency, latency, and cost without being told to — and you can point to real examples where you built, scaled, or seriously worked through large systems.
- You've shipped something real and can defend every decision — a project, open source, research, a hackathon — and go deep on why you designed it the way you did.
- You write solid code in at least one language and learn new ones fast. Our stack is mostly Python (backend and AI) and TypeScript/React (product) — you need one and the ability to pick up the other.
- Genuine curiosity about LLMs and agents. You want to build with them, not just use them.
- You move fast, own your work, and can work in person in San Francisco.
Nice-to-have — bonus, not required:
- Strong OSINT experience is a major plus — you’ve done deep online investigations, identity resolution, reverse username research, entity mapping, or similar open-source intelligence work.
- You've built something with LLMs or agents — a research tool, a RAG app, a scraper, an agent loop.
- Experience with distributed systems, queues, workflow engines, or high-throughput pipelines.
- Some React or Next.js, or experience building data-heavy UIs.
- You've run systems in production — databases (SQL/Postgres), Redis, search, observability.
- A sharp eye for data quality — you no