Nate Lamarche — Resume

Nate Lamarche GitHub | LinkedIn | natealamarche@gmail.com Education University of Waterloo Sep 2025 - Apr 2030 Bachelor of Software Engineering (BSE) Cumulative Average: 93% Relevant Coursework: Linear Algebra (100%), Calculus II & Multivariable Calculus, Data Abstraction & Implementation, Digital Circuits & Systems; In progress: Compilers, Statistics Experience WAT.ai Sep 2026 - Present Machine Learning Researcher Waterloo, Ontario • Developing reinforcement learning controllers for locomotion of a tendon-driven soft quadruped robot using PyTorch • Implementing and evaluating model-free and model-based reinforcement learning approaches for gait control • Supporting sim-to-real deployment by integrating learned control policies with physical robot hardware Ernst & Young (EY) May 2026 - Aug 2026 Software Engineering Intern Toronto, Ontario • Developed and maintained production Java/Spring Boot services for a financial institution serving 13M+ clients • Improved production resilience with retry logic, exception handling, and automated tests for Spring Boot services • Built a cloud-native banking API with AWS Lambda, SQS, DynamoDB, RDS, and Node.js for 6 unique endpoints • Implemented JWT authentication, transfer idempotency, secure file access and configured VPC networking Uplus Education Feb 2024 - Jun 2024 Web Developer Intern Remote • Implemented client-side JavaScript features for custom interactions, dynamic UI behaviour, and application logic • Analyzed website usage data to identify patterns and drive data-informed improvements to site structure and UX Projects RL Inverted Pendulum Aug 2026 - Sep 2026 Python· PyTorch· Gymnasium· NumPy • Built a sim-to-real reinforcement learning control system to balance a physical rotary inverted pendulum • Implemented PPO from scratch in PyTorch with clipped policy updates, GAE, and entropy regularization • Pretrained via Imitation Learning from an LQR controller, achieving 0.41° angular RMSE on local balancing trials • Deployed the RL policy to hardware, integrating sensor-based state estimation and closed-loop motor control Neural Chess Engine Mar 2026 - Aug 2026 C++· Python· PyTorch • Built a chess engine with 7.7M+ move-generation nodes/s and an Efficiently Updatable Neural Network (NNUE) • Designed and trained the NNUE chess evaluator in PyTorch with 45K sparse features on 50M labeled positions • Deployed 66MB model in native C++ with no external ML runtime, matching PyTorch within 1e-6 absolute error • Implemented incremental NNUE accumulator updates, improving neural search throughput by 24% to 761K nodes/s Goose Trials Dec 2025 - Jan 2026 TypeScript· React· Next.js • Led the development of a Next.js/React web app with 7 cognitive mini-games for university students • Reached 8,000+ games played, 1,000+ unique users and 100+ authenticated sign-ups within first 24 hours of launch Skills • Languages: Python, C++, C, TypeScript, JavaScript, Java, SQL • Frameworks & Libraries: PyTorch, NumPy, Pandas, Matplotlib, Gymnasium, Spring Boot, React, Next.js • Cloud & Infra: AWS (Lambda, SQS, API Gateway, Cognito, DynamoDB, RDS, S3, MSK), PostgreSQL, Git/GitHub • Technical Areas: Machine Learning, Reinforcement Learning, Backend/API Development, Cloud Infrastructure