Brahim Redouane Mellah

Senior Software Developer in Montréal: production C++ for machine vision, machine learning and deep learning

As of October 2026

Experience

Aug 2026 — Present

Zebra Technologies

Senior Software Developer, Generative AI

Agents and retrieval

I build agentic workflows and retrieval-augmented generation (RAG), and the evaluations that decide when an agent is ready to ship.

Jul 2024 — Aug 2026

Zebra Technologies

Senior Software Developer, Aurora Imaging Library

2D and 3D vision algorithms

I owned Edge Finder, Geometric Matcher and Calibration in the Aurora Imaging Library, Zebra’s machine vision SDK, from feature to release: API design, bug fixes, documentation, code examples and customer support. I agreed API behaviour and backward compatibility with the UI, documentation and QA teams, and helped plan the roadmap.

Smart cameras

I brought the library’s image processing to Zebra’s Aurora Focus smart cameras, within their limited memory and processing power. From the user interface to the backend, I fixed multithreading and concurrency bugs and built features that make the camera easier to configure.

Machine learning in C++

I implemented decision trees, random forests and k-means in the library’s C++: multithreaded, thread-safe, with early stopping.

Deep learning

I benchmarked deep learning models for object detection, OCR and text detection on CPU and GPU, tuned their hyperparameters and designed the metrics that decide whether a model ships, then took the accepted ones to production.

Jun 2022 — Jun 2024

Zebra Technologies

Software Developer II, Image Processing

The Advanced Geometric Matcher

I took the Advanced Geometric Matcher (AGM), a new pattern-matching tool, from research and prototype to production, and designed its public C++ API with the team. It ships today in Zebra’s Aurora Imaging Library.

Machine learning in geometric tools

I added random forests and SVMs to modules that had been purely geometric, so they could classify patterns in cluttered scenes, with new APIs, documentation and examples. I also mentored interns and junior developers on modern C++, multithreading and code review.

Numerical bugs and regression tests

I fixed numerical bugs and API regressions, and added unit, integration and randomised tests to the regression suites.

Jun 2019 — Jun 2022

Matrox Imaging

Software Developer, Image Processing

Matrox Imaging, in Dorval, became part of Zebra Technologies in 2022.

Calibration, measurement and metrology

I implemented and debugged the library’s Calibration, Measurement and Metrology modules, with their tests, documentation and examples, and supported the customers who used them. I also maintained Model Finder, Edge Finder and Bead.

Non-linear solvers

I replaced older geometric solvers in the library’s decades-old C++ codebase with non-linear optimisation (including Levenberg–Marquardt), for sub-pixel accuracy and numerical stability. I tracked down memory leaks and dangling references, and built regression suites that protect the behaviour thousands of installed applications depend on.

Jan 2019 — May 2019

Matrox Imaging

Software Development Intern

A random forest classifier

I built a random forest classifier into the Matrox Imaging Library (today Zebra’s Aurora Imaging Library), with multithreaded training and warm start, so production lines could sort defects in real time, and the synthetic datasets and automated benchmarks that measured its precision and speed on every change.

2017 — 2019

Polytechnique Montréal

Master of Engineering (M.Eng.), Electrical Engineering — Control Systems

Professional master’s with an industrial internship · GPA 3.87/4 · Al Ghurair Foundation Scholarship — 1 of about 100 recipients from more than 15,000 applicants.

How I got to Canada

The scholarship was awarded on academic merit and paid for the master’s. It’s how I came to Montréal.

Projects on real hardware

Every major course ended in a project on real hardware: real-time scheduling under QNX, a MATLAB and C++ vision system that located parts for a Fanuc 6-axis robot, SIFT panorama stitching, and formation tracking on the Georgia Tech Robotarium’s robot fleet.

2012 — 2017

École Nationale Polytechnique, Algiers

Engineering degree (Diplôme d’ingénieur d’État), Electrical Engineering — Control Systems

GPA 17.5/20 · ranked 5th of 1,400 in the national entrance exam after the preparatory classes (top 1%).

Preparatory classes

Two years of intensive mathematics, physics and programming, before the national ranking that decides which engineering school you enter.

Control systems on real hardware

I built an adaptive cruise controller for an autonomous vehicle and a PLC program that drives an industrial assembly machine.

Skills

Languages
C++17 / C++20  C  Python  TypeScript  Dart / Flutter  MATLAB  SQL
Machine vision
Aurora Imaging Library  OpenCV  pattern matching  calibration  metrology  edge measurement  SIFT  homography  2D and 3D geometry
Machine learning
decision trees  random forests  SVMs  k-means  scikit-learn  deep learning  object detection  OCR  model evaluation  hyperparameter tuning
Maths
non-linear optimisation  Levenberg–Marquardt  linear algebra  numerical stability  control theory
C++ & systems
modern C++  multithreading and concurrency  virtual memory  CMake  GTest  ASan / MSan  Valgrind  gdb  Windows and Linux  Docker
Generative AI
agentic workflows  RAG  agent evaluation
Tools
Git / GitHub  SVN  CI/CD  code review  clang-tidy  SonarQube  Visual Studio  JIRA  Confluence  Google Cloud  Firebase  FastAPI
Spoken
French (native)  English (professional)  Arabic (native)

Certifications

  • 2023 Neural Networks and Deep Learning · Convolutional Neural Networks DeepLearning.AI
  • 2023 Machine Learning: Classification University of Washington