Apptronik Apollo · simulated in MuJoCo

Abbas Syed

Robotics & ML Engineer

About me

I'm an engineer based out of Waterloo and Toronto, focused on physical AI and autonomous robotics. Over the past few years I've worked on robots from competition bots to humanoids, across everything from software and hardware to autonomy and AI. Right now I'm building humanoid robots with WATonomous.

What motivates me most is seeing engineering make a positive impact on people's lives, especially in medtech. I previously worked on AI/ML for medical imaging at Ramsoft. I also love photography, sports, and reading.

Experience

WATonomous

Robotics Engineer · Humanoid Team

Building UWaterloo's first humanoid robot, working across the stack on physical AI, perception, and firmware. Developing 3D SLAM and learned policies for autonomy, including RL, IL, and VLA finetuning. Working from simulation through to deployment on real hardware.

  • ROS 2
  • C++
  • PyTorch
  • Isaac Lab
  • MuJoCo
  • RTAB-Map
  • Foxglove
  • VLA finetuning
Render of the WATonomous dual-arm humanoid upper body
3D-printed humanoid hand with tendon-driven fingers and a control board

Ramsoft

AI/ML Software Engineer Intern

Worked on software and AI for medical imaging technology used by healthcare professionals in hospitals and imaging centres around the world, including a production AI assistant and a data analysis agent for the OmegaAI platform.

  • AI/ML
  • LLM agents
  • RAG
  • Azure
OmegaAI imaging viewer with worklist and radial tool menu
OmegaAI viewer showing CT measurements across a series

Computer Vision for Smart Systems Lab

Research Assistant

Developed and integrated a low-cost motion capture system for tracking autonomous drones indoors, as part of the CViSS lab at the University of Waterloo.

  • Ceres Solver
  • OpenCV
  • Python
  • Hardware

FIRST Robotics Team 2702

Software Developer & Team Lead

Built robots to compete in the FIRST Robotics Competition, working across computer vision systems, mechanical prototyping, and full-stack development. Directed team strategy by using data to select playoff alliances, directly contributing to playoff success.

  • OpenCV
  • PyTorch
  • Raspberry Pi
  • TypeScript
  • React
Team 2702's robot on the field at a FIRST Robotics competition
Team 2702's robot beside the scoring structure during a match

Project Astraeus

Co-Founder & Chair

Empowering youth to address impactful problems while learning engineering, with branches across Canada reaching over 2,000 people, supported by UWaterloo's startup incubator. Leading an engineering team of undergraduates developing and teaching machine learning and computer vision through workshops and events.

  • Python
  • PyTorch
  • OpenCV

Projects

Voice-Triggered Defect Detection & Removal

Award winner · Toyota Innovation Challenge

A hands-free quality control system for production lines: a voice-triggered camera hard hat paired with a robotic arm that spots defective parts and pulls them off the line on its own.

  • Python
  • OpenCV
  • Camera calibration
  • Speech recognition

StrobeShield ↗

Best Hardware Hack · EurekaHacks

Glasses that detect dangerous flashing light and polarize to block it before it reaches the wearer, helping prevent photosensitive seizures.

  • Raspberry Pi Pico
  • Signal processing
  • Flask
  • 3D printing
The StrobeShield prototype: 3D-printed geared glasses wired to a Raspberry Pi Pico
CAD model of the StrobeShield glasses with geared lens rings

Motion Planning: RL vs. Classical ↗

Reinforcement learning agents trained to navigate, compared against classical planners like A* and RRT in a custom-built robot physics simulator.

  • Python
  • PyTorch
  • DQN
  • A*
  • RRT