Things we've built.

A lead generation system that gathers leads via the Google Places API, then enriches them with missing fields like company domains and industries to meet CRM requirements. Other work includes Shopify storefront edits in Liquid, and a cleanup of their shipping database to reformat the schema for better storage.

An AI-agent network for drug repurposing and cancer drug discovery. The system identifies protein targets and simulates molecular docking to predict binding poses, helping researchers explore treatment options faster.

A computer vision interface that translates eye blinks into Morse code, then into text. It gives patients with Locked-in Syndrome (LiS) a way to communicate using only their eyes.

An image-analysis tool that helps businesses reduce food waste. Staff scan produce to get freshness scores and expected shelf-life, so they can prioritize stock and cut spoilage before it hits the bin.

PulseMind takes data from your wearable device and turns it into health guidance. An AI coach calls you with insights, check-ins, and recommendations based on your live biometrics.

Uses an LSTM to generate a one-day-ahead trading signal for SPY, then backtests it against buy-and-hold. The base strategy is a moving-average crossover that racks up many small losing trades in choppy markets. Smoothing the signal with the model cut trade count roughly in half and doubled the strategy's return versus the raw signal. It still trailed plain buy-and-hold over the bull-market window tested.

A React web dashboard for patient-generated health data. It shows metrics like blood pressure, glucose, sleep, activity, and symptoms through interactive charts, with separate views for patients, physicians, and administrators. A Python script generates synthetic patient data to populate the dashboard for testing and demos.