Founding ML Engineer, Forecasting

Role overview Build the forecasting engine behind a cash-flow product for individuals and freelancers. As an early engineer, you will shape how transaction histories become projections, confidence ranges, and warnings that people can use to plan ahead; the work emphasizes measured accuracy and honest uncertainty.

Responsibilities - Develop methods for detecting recurring transactions and projecting future cash flow, volatility, and confidence ranges. - Create an evaluation set of transaction histories with known outcomes and use it to assess changes before release. - Make forecasts appropriately cautious when transaction history is limited or income is irregular. - Improve ingestion so transaction data can be handled across varied bank-exported CSV formats. - Build and refine dependable systems that users can consult when making near-term financial decisions.

Requirements - At least five years building production systems that others rely on, including hands-on time-series or forecasting work. - Practical experience evaluating predictive systems and the judgment to identify weaknesses in evaluation methods. - Comfortable writing TypeScript, or able to become productive in it quickly. - Care for the difference between a convincing demonstration and a reliable product used in real decisions.

Benefits and work setup - Full-time role, based in San Francisco or remote within the United States. - Compensation listed as $170,000–$210,000 plus 0.4%–0.9% equity. - Hiring process includes an introductory call, a paid three-to-four-hour work session, and two conversations with prospective colleagues; it typically takes about two weeks.