Strobe is building the world's largest power plant: not a single site, but a distributed fleet of buildings, batteries, EVs, and generators that buy and sell power in real time across wholesale energy markets.
Own the environment our quants work in: the market, weather, and site data they research on, and the tooling that takes a forecast or strategy from a notebook to production. You'll know the data better than anyone, and build the systems that make research fast and trustworthy.
What you'd own
- Clean, point-in-time datasets of wholesale prices, load, weather forecasts, and site telemetry, with revisions and publication times tracked
- The research environment: data access, feature pipelines, and the libraries quants use to explore, fit, and compare models
- Backtesting and simulation infrastructure that replays markets and asset constraints exactly as they were
- The path to production: scheduled forecast and dispatch runs, versioned models, and monitoring that alerts when inputs or outputs look wrong
- Data quality: gaps, late and revised data, timezone and DST edge cases, and reconciliation against ISO settlements
- Close work with quants and traders on what they need next, and the judgment to build it once, properly
Strong fit if you have
- Strong Python and SQL, and research or data platforms you built that others relied on
- Experience with time-series data at scale, and an instinct for lookahead bias, leakage, and survivorship in historical data
- Enough statistics and forecasting to understand what quants are building and what their tools need to get right
- Experience in trading, market data, or another domain where a data error costs real money
- The rigor to check your own numbers before someone else has to
Bonus points
- Experience in power markets (NYISO, PJM, CAISO, ERCOT): LMPs, capacity, or ancillary services
- Weather data and numerical weather prediction inputs
- Optimization tooling (LP/MILP solvers) or battery and generator dispatch
- Columnar and time-series stores (ClickHouse or similar)
Stack: Python / SQL / ClickHouse / AWS / Terraform / Grafana. AI-agent-native monorepo with deep investment in agent-enabled engineering efficiency.
Small team, high ownership, no layers.