Projects

End-to-end ML pipeline that ingests vibration sensor streams, engineers features (RMS, kurtosis, FFT bands), trains an LSTM model, and serves predictions via a REST API — all containerised with Docker Compose.

A Python toolkit for batch-processing OpenFOAM output files: extracts pressure / velocity fields, generates publication-quality plots, and exports summary CSVs. Reduced analysis time from hours to minutes.

FastAPI + HTMX real-time dashboard that polls an InfluxDB time-series database and renders live charts without a heavy JS framework. Deployed on a Raspberry Pi in the lab.

This site — a FastAPI-powered portfolio with Jinja2 templates, Tailwind CSS, Alpine.js micro-interactions, and HTMX for progressive enhancement. Content fully driven by a single YAML file.