Teodor Ilie

About Me

Teodor Ilie

MSc Student at Queen's University | Full-Stack Developer at BMO

Download Resume

Hello, I'm Teodor Ilie

Welcome! I'm a Master's Student with the School of Computing at Queen's University in the QUARRG lab, and a full-stack software developer at BMO. I am currently researching Deep Reinforcement Learning solutions for autonomous vehicle navigation and control.

In my free time I enjoy kickboxing, playing guitar, and designing LEGO Technic models. Check out my coding projects, LEGO Technic designs, or just enjoy the pretty pictures 🎨

Interests

  • Reinforcement Learning
  • GenAI
  • Autonomous Control
  • Event Camera
  • Computer Vision
  • Deep Learning
  • Automated Planning
  • Explainable AI
  • Cybersecurity
  • FinTech

Education

M.Sc., Computer Science (AI Spec.)Queen's University, recipient of:
B.Sc. (Hons), Computer ScienceQueen's University

Projects

Explore a selection of my projects, ranging from a self-driving robot using Deep Reinforcement Learning for map building, Probabilistic Inference with Pacman, and a Neural Network digit classifier built from scratch. Also check out my collaborators below 🤖

Autonomous Racing Team Founder & Lead

Founder & Captain of Ingenuity Labs Racing, a team of PhD, post-doc, and MSc students competing internationally at F1Tenth/RoboRacer autonomous racing competitions.

Technologies Used

  • Control
  • Perception
  • Planning
  • AV

Gym-Khana: Autonomous Racing & Drifting Gym Simulator

Designed a Gymnasium environment for training Deep Reinforcement Learning policies to race and drift on 1/10 scale or full-size vehicles.

Technologies Used

  • Deep RL
  • Gym
  • SB3
  • Wandb
  • Vehicle dynamics

Deep RL for Active SLAM

Built an Active SLAM system for autonomous navigation using ROS2, Gazebo, SLAM Toolbox, and PIC4RL on a Clearpath Jackal robot.

Technologies Used

  • ROS2
  • Gazebo
  • SLAM
  • Deep RL
  • Occupancy grid
  • Pose graph optimization

CNN Handwriting Classifier

Designed and implemented a Convolutional Neural Network model in Python from scratch to classify handwritten digits, using the MNIST dataset.

Technologies Used

  • Python
  • Convolutional Neural Network

AI Pacman Game

Implemented probabilistic inference algorithms to enable Pacman to hunt invisible ghosts based on noisy distance readings.

Technologies Used

  • Inference
  • Python
  • AI

Monocular Visual Odometry in OpenCV

Implemented SIFT feature detection and brute-force matching to accomplish Monocular Visual Odometry (MVO)

Technologies Used

  • OpenCV
  • MVO
  • Carla
  • AdverCity

Theory alignment via regular bisimulation

Designed a program using tarski, a PDDL Python library, to automatically detect isomorphism, which was used to automate checking correctness of candidate solutions.

Technologies Used

  • PDDL
  • Python
  • tarski

OpenGL Scene Generation

Designed a 3D scene with ray tracing, reflection, refraction, and surface rendering in OpenGL using C++

Technologies Used

  • Computer Graphics
  • Ray Tracing
  • OpenGL

LEGO Technic Creations

My passion for LEGO Technic began over a decade ago and continues to this day. Check out an assortment of custom-built gearboxes, transmissions, suspensions, and axle designs - all engineered from the ground up. I designed every project meticulously from original blueprints, learning the joy of building something my own.

That same drive fuels my real-world engineering and coding accomplishments - building an exceptional coding project is not so different from engineering a really cool truck. P.S. Kudos to my wife for not only putting up with my obsession but helping shoot much of the footage 🫶

Builders Who Inspire Me

For setting the bar high, and showing all us builders what LEGO Technic can do, here are some notable inspirations. Thanks for the motivation, and sometimes even collaboration 🤟

Connect With Me

Send me a message