cv
A detailed resume.
Basics
Name | Qasim Ali |
Label | Research Engineer |
muhammad.qasim.ali06@gmail.com | |
Phone | +1-647-807-7243 |
Url | https://qasim-ali0.github.io |
Summary | Graduate student in AI and Robotics |
Work
-
2021.05 - 2022.08 Software Engineering Intern
Magna Electronics
Contributed to embedded software development for advanced self-driving features.
- Debugged complex detection algorithms and state machines in simulation and real-time.
- Developed communication interfaces for vehicle signal integration.
- Implemented end-to-end AWS pipeline for cloud-based KPI tracking.
- Built a tool using Unreal Engine for visualizing object and lane detection.
Education
Awards
- 2023.09.01
Engineering Excellence MASc Fellowship
University of Waterloo
$50,000 awarded for graduate research excellence.
Publications
-
2025.06.15 MR-CRL: Leveraging Predictive Representations for Contrastive Goal-Conditioned Reinforcement
RLC 2025 Workshop on Reinforcement Learning Beyond Rewards
Developed a predictive representation algorithm for goal-conditioned reinforcement learning in maze and manipulation tasks.
-
2025.05.20 Humanoid World Models: Open World Foundation Models for Humanoid Robotics
ICML 2025 Workshop on Building Physically-Plausible World Models
Trained video-based world models for humanoid robotics using diffusion, flow matching, and masked transformers. Used parameter reduction strategies to reduce model size.
-
2025.02.19 RepViT-MedSAM: Efficient Segment Anything in the Medical Images
Springer, Cham (Lecture Notes in Computer Science, vol 15458)
RepViT-MedSAM proposes an efficient, distilled alternative to MedSAM for medical image segmentation, achieving strong Dice and NSD scores while reducing computational cost. Designed as part of a Medical Imaging contest.
-
2025.01.31 GraphPad: Inference-Time 3D Scene Graph Updates for Embodied Question Answering
CVPR 2025 Workshop on Bridging Language, Vision and Action in 3D Environments
Proposed a scene graph and API-based reasoning pipeline that improved EQA benchmarks using fewer images.
-
2024.09.19 MGSO: Monocular Real-time Photometric SLAM with Efficient 3D Gaussian Splatting
ICRA 2025, Main conference
Helped develop a real-time 3D reconstruction algorithm using Gaussian Splatting and DSO SLAM. Team achieved state of the art visual quality and FPS. Note: I was not the lead author on this paper, but contributed to the gaussian splatting improvements.
Skills
Programming | |
Python | |
C | |
C++ | |
Pytorch | |
Jax |
Embedded Systems | |
ARM Cortex | |
Visual Studio | |
Canoe | |
Trace32 | |
Verilog |
Languages
English | |
Native speaker |
Interests
Robotics and AI | |
World Models | |
Generative AI | |
Reinforcement Learning | |
Humanoid Robotics | |
Representation Learning for Reinforcement Learning | |
Contrastive Reinforcement Learning | |
Model-based Reinforcement Learning | |
Embodied Reasoning | |
Video Prediction |
Projects
- 2022.09 - 2023.04
Parking Lot Drone
Autonomous drone system for detecting parking occupancy. Developed as undergrad capstone project.
- Flight videos demonstrating autonomous survey
- 2021.10 - 2023.02
Autonomous Vehicle Control
Designed autonomous control logic for V2I intersections as part of McMaster EcoCar team.
- Presented at vehicle control contest