hacker

John Chen

News

September 2023: I have made the decision to return to academia and am actively searching for a PhD position! I am primarily focused on fairness in multimodal foundational models for clinical medicine. I am also interested in general training dynamics of multimodal LLMs as well as their applied use cases (debiasing, ECG classification, etc.). Please e-mail me at johnc dot cs dot toronto dot edu to collaborate!

About

Hi! My name is John Chen and I am a second year medical student at McGill University, currently on leave. Previously, I completed my Master's and Bachelor's degree in Computer Science at University of Toronto, St. George Campus and Vector Institute under the supervision of Frank Rudzicz. During my time at U of T, I focused on a variety of topics in NLP, speech and fairness. Here are some of my selected publications:

Projects

Cllm-rnn code

Character Level Language Model

Implemented vanilla recurrent neural network to perform language modelling and text classification. Coded the entire project end-to-end, implementing the DataSet, architecture, training loop and evaluation (using Pytorch). Modified network as necessary to deal with gradient exploding (implemented manual gradient clipping) and conversion between unsupervised and supervised use cases.

Card image cap

SPANHacks Data Analysis

Performed logistic regression on a dataset provided by the Canadian Basketball Association and delivered data insights.

Card image cap

Talk to Plant

Architected and implemented the hardware system for an E2E house-plant IoT VoIP monitoring system. Talk to Plant allows you to call in from a mobile device to check on the status of your household plant, as monitored by various sensors hooked up to an ESP8266 Micro-controller.

Github

Hackathons

I've been lucky enough to attend the following hackathons! I am a big fan of hackathons as I feel they are great opportunities to sit down and build not only software, but build connections too!

  • Built Reputable @ BlockHack @ YorkU
  • Built CongoToken @ CUTC Deloitte Case Competition (Multuiparty Aggregation on the Blockchain) @ Kitchener
  • Built DataX @ ETHUofT @ U of T
  • Built StudyTracker @ CSSU X ConnectedLab Hack Night @ U of T
  • Top 50% in Optapalooza @ U of T
  • Built this site @ U of T Hacks
  • Built Project Mom @ Hack the North @ Waterloo
  • Built Talk to Plant @ Cisco DevNet Hackathon @ Toronto

Visitor Map

My current work involves a lot of the .NET stack: C#, Powershell, ASP.NET and so forth, with a parallel focus on DevOps: using and building developer tools for Github, Jenkins, Nexus and UrbanCode. Also, most likely the ebst way to make the background is to just apply a filter/blur to a good image