I design pipelines and models that stay correct under late data, scale, and real-world failure.
Currently working on audit analytics, agentic backends, and production forecasting pipelines.
linkedin github gitlab kaggle stackoverflow
- Designing ingestion and modeling systems for messy, high-volume event data
- Production ML and LLM workflows with evaluation, monitoring, and deployment hygiene
- Resilient integrations handling rate limits, backfills, schema drift, and retries
- Building a Google Workspace audit analytics pipeline with overlap-safe ingestion
- Developing agentic backend workflows using LLMs
- Writing about real-world data failures and system design tradeoffs
- LLM-powered research tooling for AI Teaching companion combining retrieval systems, experimentation workflows, and backend services
- Geospatial satellite data pipelines surfacing mineral exploration signals from noisy remote sensing data
- Near-real-time energy forecasting pipelines spanning SCADA + weather data ingestion, data warehousing, and ML-driven grid balancing
- Semantic search and recommender system and data products powering infrastructure intelligence and risk assessment for institutional investments
- OCR-driven clinical data processing pipelines transforming unstructured medical documents into usable datasets
Pinned repositories below reflect the work above.
python numpy pandas plotly folium scikit-learn tensorflow opencv pytorch fastapi selenium javascript mysql postgresql mongodb elasticsearch kibana git docker kubernetes terraform github-actions google-cloud-platform amazon-aws
Open to Data Engineering, MLOps, and Platform roles. Best reached via LinkedIn or email.
This README is generated every 24 hours!
Last refresh: 02:10:32 GMT+0000 (Coordinated Universal Time)