Jessee Lord

Created a local startup providing fall detection solutions for elderly care. Developed and integrated a C++ and Python program into CCTV systems to detect and alert in case of falls. Built an admin dashboard for real-time monitoring, with all data stored securely on client-owned servers. The system is deployed for retirement homes and can also be installed in private residences.

As a Level Designer at Ubisoft GameLab, I worked with a multidisciplinary team to design, prototype, and balance immersive game levels. I contributed to creating interactive experiences in Unity, integrating new gameplay mechanics and ensuring a smooth game flow. My role involved iterating quickly based on player feedback, optimizing environments for performance, and clearly documenting design intentions to keep communication seamless between artists, programmers, and game designers.

Developed a robotic firefighting system using an infrared camera to detect heat sources and trigger an autonomous response. Designed and programmed a robotic vehicle equipped with a water pump to navigate toward the fire and extinguish it. Integrated real-time detection with automated navigation for rapid incident response, combining computer vision and robotics in a functional prototype.

Documents the six-stage evolution of the OpenMTL bus-position correction algorithm — from instant GPS snapping to a horizon-aware additive controller. The final stage spreads tracking error across the entire remaining route using a precomputed correction rate, achieving smooth segment-invariant convergence at O(1) per frame with Lyapunov-stable error dynamics.

Trains a GPT-style transformer (~15M parameters) on 31M tokens of STM GTFS data to learn the structural grammar of Montreal's bus and metro network. Achieves 99.4% next-token prediction accuracy and 98.8% next-stop accuracy. Includes route embedding clustering, network bottleneck identification via coefficient of variation analysis, schedule padding detection, and route variant discovery.