My time at Romeo Power as a Software Engineer, starting in July 2021, was one of the most intense and rewarding phases of my career. The electric vehicle industry was booming, but with its rapid growth came alarming news—reports of EV battery fires were becoming more frequent. Every few weeks, another case would surface, often with catastrophic consequences.
Within our team, this wasn't just another problem to solve—it was a matter of safety, of lives at stake. We took it personally. We knew that if we could develop a way to predict these failures before they happened, we could make a real difference.
For months, we analyzed countless fire incidents, pouring over battery telemetry data, failure reports, and real-world case studies. The challenge was immense. Unlike mechanical failures, which often show clear warning signs, thermal runaway in lithium-ion batteries can escalate rapidly, giving users little time to react. The key was to find subtle indicators—minute fluctuations in voltage, temperature, or current—that could hint at an impending failure before it was too late.
After many discussions, experiments, and dead-end approaches, we finally devised an innovative algorithm. The breakthrough came when we identified a combination of data points that, when monitored together, provided an early warning—just minutes before thermal runaway occurred. It wasn't much time, but it was enough. Enough for an alert to be sent, for a vehicle to be powered down, for a user to exit safely.
Once we had the algorithm, the real work began. I was responsible for implementing it, ensuring it could run in real-time without adding significant computational overhead. Testing was rigorous. We simulated various failure scenarios, analyzing how our model responded. The first time we saw it successfully predict a thermal event in a test environment was exhilarating. We knew we had something that could work.
Finally, after months of validation, the algorithm was deployed. We integrated it into the battery management system, allowing vehicles to monitor themselves continuously and alert users if danger was imminent.
Years later, as I reflect on this, I wonder how many lives were quietly saved because of our work. We may never know the exact number, but knowing that we helped make electric vehicles safer is enough. That's the beauty of engineering—sometimes, the greatest impact comes from the things people never even realize were protecting them.