TremOR · Founder
Low-cost sensors that listen to buildings
Funded through ShahLab.AI's AI for Social Good Competition, 2026
After an earthquake, engineers have to check thousands of buildings fast, and much of the damage in brick and masonry is invisible. TremOR turns inexpensive automotive knock sensors into a vibration-reading tool: tap a wall, record how it rings, and let a model flag hidden cracks, mortar separation or detached plaster for an engineer to check first.
Because the hardware is cheap, the same approach could serve post-disaster surveys, routine inspection of older buildings and industrial condition monitoring. Structural engineers at Miyamoto International and masonry researchers have given feedback on the method, and the ShahLab.AI judges described TremOR as relevant, timely and promising for impact.
signal processingembedded sensingmachine learningstructural engineering
Tap response, illustrated: healthy brick (blue) vs. hidden crack (coral)
Purdue AI learning initiative · Student lead
Helping students use AI well
I lead a student effort to help Purdue students use AI more effectively and more responsibly, so they learn more now and are ready for a workforce that keeps changing. Instead of adding yet another tool, the idea is to connect what Purdue already has: course-grounded AI assistants that follow each professor's materials and live inside Brightspace.
That has meant conversations with deans, Purdue's instructional technology team and Google about campus-wide access, course integrations and a virtual-TA pilot.
AI in educationBrightspaceGeminiresponsible AI
Steady Flow · Founder
A study companion for engineers
A learning app that starts from each course's own notes and instructor methods, separates "I've seen this" from "I can do this," and plans review around real exam dates.
Entalvo · Founder
Public bids, found for you
A web service that watches U.S. public-procurement sources and surfaces RFPs, RFQs and RFIs for small construction firms, reading public sites only and following each site's rules.
Independent research
Teaching AI to keep learning
I'm researching continual learning: how models can pick up new skills without forgetting old ones, including ideas borrowed from neuroscience. I run it with a small team of AI models from different providers that propose, run and review experiments, while the test data stays hidden from them.
ML researchmulti-agent systemsevaluation
How the research team is organized: one lead, specialist agents, independent review