Lead Modelling Software Engineer

About Homely

At Homely, we’re dedicated to shaping a zero-carbon future. Our smart technology makes it easy for both installers and customers to transition to heat pumps by providing industry-leading heating controls. By predicting and optimising using unique learned house characteristics, time-of-use electricity tariffs, weather forecasts, and solar PV data, we lower energy bills and reduce CO2 emissions.

The Role

We’re looking for someone who is ready to take their next step and join us as Lead Modelling Software Engineer to own, scale, and maintain the core predictive and optimisation engines powering our smart controller. This isn't a standard CRUD-app software role: you will be translating the actual physics of home heating and weather forecasting into robust, production-ready code which serves thousands of customers. You'll join a close-knit team where you'll own the full lifecycle of our algorithms, taking ideas from initial research and simulation, building the MVP, and deploying them to our datacentre and on-device ecosystem.

What you'll be doing day-to-day

• Model Production & Optimisation: Translate mathematical formulas, control theory, and thermal dynamics into clean, efficient, and maintainable code.
• Taming Real-World Data: Maintain and improve our parameter-finding suite. You’ll figure out how to clean, filter, and make sense of noisy, fragmented telemetry from physical sensors to accurately map out how individual houses heat up and cool down.
• End-to-End Ownership: Take features from a notebook or simulation to an MVP deployment, building code that doesn't break when hit with real-world IoT anomalies.
• Testing & Validation: Build robust testing suites and simulation environments to stress-test optimisation logic against historical tariff, smart meter, and weather datasets.
• Code Quality & Teamwork: Keep our standards high. Participate in code reviews, enforce strict data validation, and collaborate across teams to ensure models integrate smoothly with our production pipelines.
• Keeping Things Running: Keep an eye on how your models perform in the wild, proactively troubleshooting alerts, bugs, and quirky edge cases to keep the system stable.