
FSDAD - Generalizing Deepfake Audio Detection
First time trying research, wrote a solo paper on framing deepfake audio detection as a meta-learning problem to try few-shot generalization of a deepfake audio detector to new deepfake generators.
Learning about ML
Prev. @ Shopify, 2nd Year Computer Science @ University of Toronto, Schulich Leader.
ML and ML infra for Messaging team, smart sending, impact projections, causal learning.
Engineering and research for UofT Machine Intelligence Student Team's FixMyElo, DFOD, and Agent Forge projects

First time trying research, wrote a solo paper on framing deepfake audio detection as a meta-learning problem to try few-shot generalization of a deepfake audio detector to new deepfake generators.

Architected a self-explaining RL-based chess agent by using attention-weighted board states, policy/value networks and MCTS + UCT move calculation with PyTorch, CUDA and python-chess

Built an encoder-only transformer from scratch + discord bot for 11th grade CS class. Determines if school board Instagram posts indicate snow days.

Founded a 100+ student high school hackathon. Built the website, hosted 3 workshops, and I was the keynote speaker!

Founded full-stack medical app to schedule immunization and cancer screening dates. Led entire SDLC and developed scheduling algorithms based on user-provided info.

An online coding judge built to host problems and contests for the coding club at the Dr. G.W. Williams Secondary School. Inspired by https://dmoj.ca/. My first ever project!

Wrote, filmed, directed and acted in small short film just for fun!

Weird Geometry Dash memory layout idea.

Probably the best layout I've ever created in Geometry Dash.

Hackathon project that uses 3DGS to reconstruct VR environments from plain video. I think the demo video we filmed is the cooler part though.
Thoughts on things I learn, stories, and anything else I feel like writing about! — read the blog.