Machine Learning, LLMs, and Agents for Software Analysis

· Worked on development and integration of Machine Learning models· Implemented monitoring for ML model pipelines for retraining, inference, and data validation· Implemented Full CI/CD on ML pipelines using AWSNote: Position eliminated due to mass layoffs

· Intelligent Process Automation using Reinforcement Learning· Synthetic Datasets using Generative Models· Forecasting using Deep Learning · contributed to the open-source joint venture between the Vector Institute and Telus for Model Based Reinforcement Learning for datacenter HVAC control found in "https://github.com/VectorInstitute/MBRL-HVAC-Energy-Optimization"

• Improved accuracy on filtered‑backprojection single‑shot reconstruction in photoacoustic tomography (PAT) using GPU programming (CUDA C)• Implemented state‑of‑the‑art iterative reconstruction algorithm (FISTA with a Total Variational Regularizer for PAT)• Using an image‑translation GAN, reduced Radio Frequency data artifacts and background noise

• Teaching assistant for APSC 160 - Introduction to Computation in Engineering Design• Administer weekly labs and quizzes for 60+ students• Marking student lab submissions and exams• Q&A sessions both in person and online

Involved in two major projects: Photoacoustic Imaging:• Computation tecniques ofr Photoacoustic Tomography (PAT) image reconstruction• Implementation of the Delay-and-Sum algorithm in CUDA C• Writing a pipeline for image reconstruction for the NVIDIA DGX machineDigital Pathology:• Pipeline for automatic segmentation of mice brain histology slides• Feature extraction and selection for classical machine learning algorithms

• Implementation of behavioral robotic programming in C++ including clinical user feedback interface using concepts such as state machines, inheritance, and polymorphism • Migration of complex nested Subversion software tree to a more modern Git tree structure+ Exploration and implementation of automated software tree builds