Quantitative Software Engineer

About Us

Fractal Power is a vertically integrated power company built for the modern grid. We own and operate across the entire stack, combining utility-scale energy storage infrastructure with algorithmic trading and dispatch to deliver power precisely where and when it's needed most.

We are backed by the world's leading infrastructure and deeptech investors, including Riot, Definition, and Worldbuild.

Why Join Us?

Demand for power is growing at the fastest rate since WWII yet supply is stranded behind worsening transmission and interconnection bottlenecks. The new economy requires a new model of power delivery, one that's vertically integrated, algorithmically operated, and co-located with demand.

We solve the last-mile infrastructure problem by vertically integrating across the entire value chain of American power - from greenfield development, to algorithmic optimization, dispatch, and power trading across a fleet of distributed resources - allowing us to deliver power where and when it's needed most on the grid. We have the ambition, the team, and the zeal to create the next great power company and we are looking for fanatical people to join us.

About the role

You'll own the experimentation and data platform end to end - from the research UI down to the data infrastructure, the models, and compute underneath it. This platform powers the data and strategy behind our market operations - battery optimization, DART/PTP, CRR/FTR trading - every strategy that goes live starts here. There's no platform team to hand the unglamorous half to, and no one else to point at when a backtest gives the wrong answer.

It's a lean team with a lot of responsibility per person. No politics, no layers of sign-off - just building and shipping, with the people who'll actually use what you build.

What you'll build

• Experimentation platform. Run thousands of backtests and simulations in parallel - point-in-time correct, no look-ahead leakage, fully reproducible and traceable back to the run that produced them. One framework across battery, DART, PTP, and CRR/FTR strategies. Fast and cheap at volume: sweeps across parameters, nodes, and historical periods.

• Data platform. Pipelines for market, weather, forecast, and asset data - ingested, versioned, backfilled, monitored, and quick to extend when a new source shows up. An access layer on top so quants and strategies can pull what they need without waiting on you.

• Forecast platform. Generate the forecasts strategies run on - scheduled, versioned, and monitored. A bad forecast fails silently and shows up as a bad trade.

• Strategy development support and management. Shared tooling so quants go from idea to tested hypothesis without reinventing scaffolding each time - plus the path from experimentation to live trading, with every strategy versioned and traceable from research run to live bid, and CI/CD that makes shipping routine and reversible.

What we're looking for

• 5+ years building production systems, including something you took from zero to running

• End-to-end ownership . Comfortable owning the experimentation platform from the researcher/quant-facing UI down through the data infrastructure, compute orchestration, and backtesting engines that support it

• Strong product ownership. You're building for a handful of demanding quants and researchers with no PM in between - you should be able to watch someone iterate on a strategy or model, spot what's actually slowing them down (stale data, slow backtest cycles, brittle pipelines), and design something they'll use without being asked twice

• Tight collaboration with quants and traders -gathering requirements directly from the people running experiments, and turning them around fast

• Directness, and appetite for a domain you are less familiar with. Understanding of quantitative trading, backtesting methodology, or energy markets is a plus, not a requirement - but you'll need to ramp up quickly

• Comfort with the full stack of an experimentation platform: data ingestion and storage at scale, reproducible research environments, versioning of strategies/models/data, and clear observability into what's running and why it succeeded or failed

• An appetite for intensity. This job moves fast and doesn't let up - you should genuinely enjoy that pace, not just tolerate it