Hi, I'm Filip Grigorov

Founding Lead Applied Scientist

Researching and building efficient AI systems, including large-language-model compression, pruning and sparsification, model routing, and prompt optimization. Developed methods aimed at preserving principal subspaces and model quality under aggressive compression, and built routing systems that direct requests to the most cost-effective capable model. Also use agentic coding tools to accelerate research iteration and engineering velocity.

Machine Learning Technical Lead – Calibration, Odometry & Localization

Led calibration, odometry, and localization work for autonomous driving. Owned and evolved classical and learned visual-odometry architectures, advanced keypoint-based perception models, implemented stereo co-factor-graph visual odometry, defined the technical roadmap, and mentored engineers.

Staff Computer Vision / Deep Learning Engineer

Drove deployment of perception modules on target hardware in real time after the Algolux acquisition. Led research on weather-robust perception, helped define image inference product requirements, and contributed to the design and integration of core features across the self-driving software stack.

Senior Computer Vision / Deep Learning Researcher & Engineer

Led R&D for multi-object 2D/3D tracking and stereo-based perception, improving tracking accuracy and enabling dense depth estimation. Architected and deployed real-time embedded detection models and TensorRT-accelerated networks as part of the production perception stack and inference engine.

Senior Computer Vision / Machine Learning Researcher

Built multi-face tracking, tailgate detection, and spoof-detection systems for access control. Designed auto-annotation and data pipelines, optimized models for edge devices, and integrated computer vision algorithms into a highly optimized C++/CUDA production codebase.

Computer Vision / Machine Learning Researcher

Developed intelligent systems for robotic mobile-game testing, including detection pipelines, OCR, reinforcement learning prototypes, and automation frameworks. Worked across C++, Python, and Lua to integrate perception, control, and tooling into a unified testing platform.

Academic Intern Supervisor & Mentor

Supervised and mentored an intern for the Winter 2023 term, and another for the Fall 2024 and Winter 2025 terms, supporting research skills, engineering practices, and project execution.

Computer Vision Intern Supervisor

Supervised two intern research projects focused on bad-light detection and head-pose estimation from minimal data inputs, guiding algorithm design and implementation.

PhD in Mechanical and Mechatronics Engineering (Robotics)

University of Waterloo, Waterloo, ON, Canada

2026 – 2030

Graduate part-time research in the Robotic Interaction, Perception and Learning Lab (RIPL) under Professor Russell Buchanan, with a research focus in robotics.

MSc in Computer Science (Robotics)

University of Toronto, Toronto, ON, Canada

2023 – 2026

Graduate part-time research in the embARC research group under Professor Nandita Vijaykumar, focused on computer systems, computer vision, generative AI for simulation, robotics, multimodal training, and reinforcement learning.

MSc Thesis: FOCUS: From Privileged States to RGB-D with Controlled Modality Switching and Representation Alignment

Affiliations:

• Faculty Affiliate Researcher, Vector Institute
• Graduate Student Researcher, University of Toronto Robotics Institute

BEng in Mechanical Engineering (Dean’s Honours List)

McGill University, Montreal, QC, Canada

2010 – 2013

Training in mechanical design, dynamics, and control, with emphasis on applied mathematics and engineering fundamentals.

BCom in Finance (Accounting Concentration)

McGill University, Montreal, QC, Canada

2006 – 2009

Business and finance education with a concentration in accounting, providing a foundation in quantitative analysis and decision-making.

DEC in Pure Science (Dean’s Honours List)

Dawson College (Honours Program), Montreal, QC, Canada

2004 – 2006

Intensive honours program in physics and mathematics, building a strong analytical and scientific background.