Introduction

Background

• Zambian 🇿🇲 | Canadian 🇨🇦
• Passionate about building safe AI systems that are conscious of the fundamental assumptions underlying them 🏗️
• Committed to a systematic process unfolding at the grassroots around the world to build vibrant communities and a better society 🌱

Education

• Bachelor's of Software Engineering Co-op, McGill University, 2025 (3.99/4.0) 🏫
• MSc in Advanced Computer Science (Artificial Intelligence), University of Oxford, 2026 (Distinction) 🏰
  • Thesis in AI safety 🔒

Work Experience

Most of my software development experience is with proprietary, closed-source production code that I cannot share. This was over the course of 4 internships in Industry:

• ML Intern @ Autodesk
  • Demonstrated technical leadership on a team of 4 to deliver from the ground up a secure, scalable, and centralized cloud platform that supports MCP servers and hosts and runs customized LLMs and agents on demand
• Full Stack Developer Intern @ Autodesk
  • Designed and implemented a full-stack feature for monitoring background jobs within Flow Production Tracking, an engineering Emmy and Sci-Tech Oscar award-winning software used by major movie and game design studios worldwide
• Software Developer Intern @ Coveo
  • Completely redesigned and implemented the object-oriented, multithreaded + multi-process architecture for a large testing bot in Python, nearly doubling its size, and deploying it to the cloud with a 60% slimmer docker image in a Kubernetes container
• Co-op Software Developer @ Teranet
  • Heavily implicated in architectural and technical upgrades to source and test code for both business methods and security systems of 3 large legal applications

Research Projects

• Existing Defenses Against Distillation Attacks Fail To Prevent Stealing Reasoning 🛡️
  • Master's Thesis at Oxford with OATML under the supervision of Yarin Gal. Code to be released soon.
  • Paper in the Technical AI Governance Research workshop @ ICML 2026
• Exploring the Effect of Geometric Information on Message Passing 📈
  • Comparing 5 of the most popular GNN architectures for molecular property prediction, and examining how their different uses of geometric information affect message passing (paper)
• Extending Deep Bayesian Active Learning with Image Data 🖼️
  • Replicating the results of work with the same title (see here), clarifying minor errors, and running additional experiments (paper)
• ImagiNeuron - Machine Intelligence for Neuron Detection (MIND) 📷
  • An 8-month research project to develop a cutting-edge approach to automatically label and visualize 3D high-resolution images of neurons in the brains of rats. We are preparing a paper for submission to IEEE Transactions on Image Processing, entitled "A Platform for the Semi-automatic Segmentation, Characteristic Shape Analysis and Simulation of Neuronal Somas"
• Applied Machine Learning 🤖 - A collection of 4 machine learning projects and papers, with high-quality writing and results (average grade of 97%)
• Bias in Natural Language Inference Systems 📘
  • A project investigating gender bias in Natural Language Inference systems in a more nuanced and novel way (see paper)

Software Projects

• Purposeful 🤔
  • An application to encourage students to collaborate on ideas that are purposeful, and address needs in their community
• Mesmorize 🧠
  • An application to help people with memorization. Built in response to a need I saw in my community
• Operating Systems ⚙️
  • Three projects in C, including multi-processing and pipeing to build a custom terminal, multi-processing to create a task scheduler, and designing and building a custom file system. Did not receive permission to release code
• SearchScripting 🔍
  • An application to streamline the search through messages of the Universal House of Justice (see here). Built in response to a need I saw in my community. GitHub repository coming soon!
• Fire Fighting Robot 🔥
  • Python code to control a robot built with the Brick Pi with Lego, Lego motors, and Lego color sensors
• Software Validation 🏁
  • Spring Boot Application to write unit tests, functional tests, and performance tests for a dummy application (average grade of 100%)
• Circuit Simulation 🔌
  • An approach using MatLab and Monte Carlo simulation and polynomial chaos to simulate the behavior of a simple circuit

Open Source Libraries and Contributions

During my recent internship at Autodesk in Machine Learning, I worked closely with agentic workflows and MCP servers. I closely followed the development of Amazon Strands from the day it was released, and contributed suggestions and feedback to the way it orchestrates multi-agent workflows