Alex Steele
Simulation Software Engineer
Tesla
About
Hello! I’m Alex, a software engineer specializing in embedded and systems programming.
I currently work at Tesla, where I optimize the distributed vehicle simulations that validate the safety of all software rolled out to vehicles. My main professional interest is designing high-performance systems responsible for mission-critical decisions. More concretely, I enjoy working on embedded firmware, high-performance applications, and AI (specifically reinforcement learning) systems. In previous roles, I’ve written safety-critical firmware for responding to automotive crashes, magnetohydrodynamic simulation software, and a satellite firmware image deployment system.
I received my Honours Bachelor of Science in Computer Science and Physics from the University of British Columbia in May 2024. In my final year, I completed my thesis on reinforcement learning for snake-like robots under the supervision of Professor Michiel van de Panne. I was also the lead teaching assistant for CPSC 110, the university’s major-stream introductory computer science course, for four years. As part of that role, I developed a high-performance code similarity detection system capable of comparing tens of millions of pairs of assignment submissions, which is now in use across the department.
Please don’t hesitate to reach out if you’d like to chat. The best way to get ahold of me is via email.
- Embedded Software
- Reinforcement Learning
- Robotic Motion Control
BSc in Computer Science and Physics, 2024
University of British Columbia
Experience
- Developed C firmware for communication over the vehicle’s controller area network, enabling full archival of crash data.
- Independently discovered and patched allocation inefficiency in Rust simulation codebase, improving runtimes by 30%.
- Optimized Rust data intake pipeline by hand-writing a parser, resulting in an 80% performance improvement.
- Wrote Rust software-in-the-loop (SIL) models and integration tests for inter-chip SPI communication and snooping.
- Ported internal hardware-in-the-loop (HIL) testing library to Linux and enabled firmware engineers to remotely execute tests using Python, C, Docker, and Jenkins.
- Designed and wrote SPI drivers to control the restraint control module’s (RCM) inertial measurement units (IMUs) in C.
- Implemented numerical integral approximations for the RCM crash algorithm’s near-deploy calculations in C.
- Created chip-level software-in-the-loop (SIL) models for the RCM’s onboard IMUs with extensive fault-injection capabilities in Rust and PyO3 and wrote SIL tests for drivers and crash algorithm using PyTest.
- Reduced hardware-in-the-loop (HIL) test execution time from 5.5 hours to 2 minutes and enabled the addition of the HIL test suite to continuous integration (CI) by automating test running using Python and C.
- Created novel discretization technique for solving ill-conditioned instances of the Helmholtz equation in MATLAB.
- Developed high-performance finite-element magnetohydrodynamic simulation software using C++ and Eigen.
- Optimized the performance of simulations with millions of degrees of freedom using knowledge of vector calculus.
- Architected and created drivers and a multithreaded Python application for the display and keypad on Kepler’s next-generation modems.
- Singlehandedly developed new remote software image deployment system capable of supporting the growing number of models in Kepler’s constellation of 19 satellites using Python and SQL.
- Maintained Racket autograder server used by over 800 students to submit and receive feedback on over 1500 files daily.
- Improved students’ engagement by providing personalized feedback using applications developed in Python and Bash.
- Detected over 200 cases of academic misconduct by designing, implementing, and deploying novel code-similarity algorithm using Rust, Python, and TensorFlow.
- Supervised three other teaching assistants who contribute to the course infrastructure and teaching materials.
