Doha, Qatar · Iberdrola Innovation Middle East
I build ML models that understand energy consumption, and the backend systems that put them into production.
About
I work on Non-Intrusive Load Monitoring (NILM) — detecting appliances like EVs, HVAC, and solar panels from smart meter data. Full ML lifecycle from feature engineering to inference.
I build APIs and services that expose ML models to the real world. From REST endpoints to Cloud Run job orchestration, I own the path from model to dashboard.
Projects
Semi-supervised pipeline to detect EV charging events from 15-minute smart meter data. Combines a house-level autoencoder with handcrafted energy features, fed into XGBoost with SMOTE for class imbalance.
Decomposing total household energy consumption into per-appliance signals using time-series feature extraction and clustering on 15-minute smart meter intervals.
End-to-end platform integrating Python ML models with a Java/Spring Boot backend. Ingests smart meter CSV data, runs predictions, and feeds results into client dashboards.
Spring Boot integration with Google Cloud Run v2 to submit, monitor, and cancel long-running ML jobs. Handles IAM configuration and full job lifecycle via the Cloud Run REST API.
Open to collaborations, questions, or just a chat.