AirAware
Forecasts NO2 and O3 levels 1 to 48 hours ahead for eight Indian metros, with calibrated uncertainty intervals and a system that declines to forecast when its inputs don't support one.
- Python
- XGBoost
- FastAPI
- Streamlit
- MLflow
- Docker

Local time, IST:
Backend Engineer | AI/ML Builder
Backend engineer. I build the APIs, data and automation that products run on.
Open to backend & AI/ML roles
I work at a startup in the market research space, on the backend side of the product: the services and APIs, the data structures underneath them, and the automation that keeps repetitive work from being done by hand. A good share of the work is integration — connecting AI capabilities and communication platforms into the product so they behave like part of the system rather than a bolted-on feature.
Forecasts NO2 and O3 levels 1 to 48 hours ahead for eight Indian metros, with calibrated uncertainty intervals and a system that declines to forecast when its inputs don't support one.
Research prototype for real-time arrhythmia detection from wearable-style ECG streams, built around a compact CNN-LSTM quantized to TensorFlow Lite for edge deployment. Not a medical device.
Selected as a facilitator for the 2026 cohort, supporting participants through hands-on Google Cloud labs and helping them get past the points where they get stuck.
Leadership role in the community: coordinating members, planning activities and keeping things running between events.
Ongoing involvement in developer communities — bringing students into backend and cloud tooling, and answering the questions that usually stop beginners early.
Organising technical sessions and events end to end: planning, logistics, speakers and follow-up.