Systems Engineer Intern
Bloom Energy
- ~5 kW MAE
- 2.5 MW recovered
- 26,000+ systems in <1 min
- Developed an XGBoost machine learning model to predict power gain after system fixes, achieving a ~5 kW MAE and outperforming a linear baseline.
- Deployed the model through a Next.js and MUI dashboard that ranked systems by predicted power gain, helping guide fixes across 183 systems that recovered 2.5 MW in one week.
- Built a Python script to retrieve and process operational data for 26,000+ systems in under one minute, replacing a 5–7 minute Excel workflow and exporting results to CSV.
- Created a power trend analysis tool to compare current and historical system output, allowing engineers to track performance changes over time.









