Battery Fleet Control & Optimization Engineer

EQORE Inc.; Somerville (MA); On-site; USD 140,000 - 200,000; Full time

Job summary

EQORE Inc. in Somerville, MA is seeking a hands-on engineer to improve QoreAI dispatch across demand charges, tariffs, and grid programs.

You’ll own major parts of the stack—from forecasts to dispatch to the tooling that proves an improvement is real—while keeping the fleet safe as it scales. You’ll build load and event forecasts, refine optimization baselines, and implement simulation and shadow-mode tooling to catch regressions before customers notice.

Qualifications

• Hands-on experience building control, optimization, or forecasting systems that ran on real data.
• Strong Python and the ability to own work from problem formulation through deployment.
• Solid grounding in optimization (LP/MILP, MPC) or reinforcement learning, ideally both.
• Experience with safe deployment and rigorous evaluation in closed-loop systems.
• An ownership mindset: you own outcomes, not tasks.

Responsibilities

• Improve QoreAI dispatch for peak shaving, ConnectedSolutions, Clean Peak, ICAP, and other programs.
• Build load and event forecasts and use them in control decisions.
• Maintain strong optimization baselines and safe fallbacks, and measure every new approach against them.
• Build simulation, backtesting, shadow-mode, and release tooling that catches regressions before they reach customers.
• Generalize controllers across sites, tariffs, operating patterns, and hardware.
• Monitor live performance, diagnose problems, and fix them.

Skills

Python

Optimization

Forecasting

Reinforcement learning

Tools

JAX

PyTorch

Job description

EQORE Inc. in Somerville, MA is seeking a hands-on engineer to improve QoreAI dispatch across demand charges, tariffs, and grid programs.

You’ll own major parts of the stack—from forecasts to dispatch to the tooling that proves an improvement is real—while keeping the fleet safe as it scales. You’ll build load and event forecasts, refine optimization baselines, and implement simulation and shadow-mode tooling to catch regressions before customers notice.