Data Lead

Description:

At most data companies, a dataset is the output of a large analyst team. Here, it's the output of a fleet of AI agents - directed by one person who stakes their reputation on it being right.

That's this role. You're not downstream. You're not cleaning data someone else built. You start from the source - raw documents, filings, internet data - and you build the dataset.

You're the reason institutional clients - banks, hedge funds, investment firms - trust the data.

AI agents do the extraction. You direct them, interrogate the output, catch what they miss and encode your judgement into validation systems that make the whole pipeline better over time.

You'll also work directly with clients, explaining methodology to sophisticated buyers who need to understand what they're relying on.

This isn't QA. It's the highest-ownership, most client-visible position in the company. Your name will be on the data. And for that, you'll have a very senior seat at the table.

✅ Must have requirements:

• Roughly 2-5 years working directly with data at a company where data is the core product - not as an analyst consuming clean datasets, but building them from messy, unstructured sources
• Demonstrable obsession with data quality - you know what great data looks like because you've spent time making it from scratch
• You go to the row level. A missing data point or unexplained anomaly bothers you until it's resolved - not flagged and forgotten
• Genuinely AI-native: you've been using agentic tooling long enough to have opinions on it - Claude Code, Cursor, OpenAI Agents SDK or equivalent, used on daily basis
• Python and SQL fluent - you've built and debugged pipelines, not just queried tables
• Comfortable in the terminal, in codebases and in real-world messy data environments