Seven projects: four built at work, three at university. Each one lists the result it was measured on.
A retrieval-augmented support assistant on AWS Bedrock, wired into the company quote system and fielding support, HR, and sales questions across the business.
Gesture-recognition models running on Raspberry Pi and Jetson Nano, turning hand movement into direct control over a sterilization rig.
Named-entity recognition over pull request diffs and discussion, automating the first pass of code review for the platform team.
A Chrome extension that read ServiceNow ticket text, surfaced the underlying issue through topic modelling, and suggested prior resolutions inline.
Object recognition and tracking driving a robotic arm and gripper, collecting tennis balls autonomously across three court surfaces.
An NLP pipeline covering translation, summarization, and classification, built on LSTM and BERT with LDA for topic structure.
Classifier comparison over US personal income data, measuring how the pandemic moved annual earnings across demographic groups.