Research Assistant — Secure cloud platforms (Permanent Part-time)
Secure delivery and research platforms on GCP: CI/CD ownership, test automation, and risk tooling for analysts.
Highlights
Refactored a monolith into per-service repos and ran 5 GitLab CI/CD pipelines plus runners on GCP across two GKE clusters (Terraform networking and observability) for repeatable cloud releases
Containerized workloads in Docker, operated research GKE clusters on GCP (workload networking), and rolled out services with Terraform so infrastructure changes were safer to ship
Added JUnit and Playwright E2E suites to 5 Java microservices to catch regressions before release
Connected Android mobile builds to GitLab CI/CD with promotion into the Terraform-managed GKE environment on GCP, keeping mobile release paths in sync with Kubernetes-backed backend delivery
Built scripted cybersecurity risk analysis for 1000+ risks across 181+ threat categories: weighted scoring, star-based prioritization, and an Angular UI so analysts could triage with consistent context
DevSecOps course delivery, GitLab fleet for many teams, and hands-on coaching on secure release practices (LOG 8100).
LOG 8100 cohort at Polytechnique MontréalGuest session with Arctiq on DevSecOps and AI-assisted secure development
Highlights
Under LOG 8100 secure-delivery goals, authored and refreshed DevSecOps-aligned lab and lecture artifacts so students practiced current release-security patterns
Across approximately 20 concurrent teams, built and maintained the GitLab workspace layout (repos, CI/CD templates, runners, workflows) with permissioning, reducing admin friction during the term
When project teams stalled on platform basics, coached container build, deploy, and debug flows in Docker and Kubernetes, unblocking lab milestones
Taught security with OWASP-driven examples (common risks, safe defaults, remediation) tied to hands-on lab scenarios
Graded labs and coached on assignments so resubmissions came back cleaner and faster
Technical Skills: GitLab, CI/CD, Docker, Kubernetes, OWASP, Linux
Soft Skills: Team coaching, constructive feedback, classroom facilitation
Labs, integrator-project support, and office hours across mobile/UX, integrator-project, and distributed-data tracks; course-level detail follows by term.
Contact hours: approximately 150 hours with students (labs, integrator-project support, office hours).
ÉTS teaching cohort (labs and integrator-project support)
Highlights
Taught labs, integrator-project support, and office hours on mobile/UX, integrator-project, and distributed-data tracks across multiple ÉTS terms, approximately 250 class contact hours total (e.g., TCH 057 Winter 2024; IND 500 Fall 2025; TCH 099 Summer 2024, Winter 2025, Summer 2025; GTI 660 Summer 2024; GTI 320 Winter 2026)
When distributed-data labs needed modernization, authored and refreshed SQL, MongoDB, Kafka, and PostgreSQL exercises with rubrics and actionable feedback for IND 500, GTI 660, and related cohorts, matching assessments to learning outcomes
For TCH 099 integrator teams blocked by integration issues, debugged architecture and backend/data pipelines, sustained Azure-backed course infrastructure, and coached milestones and SCRUM routines so students shipped end-to-end features
Facilitated TCH 057 mobile/UX labs and GTI 320 (Winter 2026) delivery, keeping sessions on track by resolving tooling and concept questions quickly
Created and updated lab content (SQL, MongoDB, Kafka, PostgreSQL) matched to modern distributed data concepts; designed and corrected lab assignments with detailed feedback and grading rubrics.
Supported students with debugging, architecture choices, and project integration (backend + data pipelines); helped instructors update/test course infrastructure across cohorts (including Microsoft Azure where relevant); guided teams on delivery practices and project execution (requirements clarification, milestones, SCRUM basics).
On the Italiarail stack with Railkey Tech: integrated Braintree, updated ticketing UI to new brand guidelines, extended purchase APIs, and tightened booking flow performance
Accelerated internal support by building Django, Pyramid, and jQuery tools plus a currency-usage dashboard, replacing ad hoc spreadsheets for CS and finance
During co-op, rebuilt the Selenium end-to-end harness and, while part-time, expanded automated coverage, catching UI regressions before release
Camera/ISP automation, vision data campaigns, and lab infrastructure for imaging R&D.
Highlights
Automated camera/ISP validation by combining Ansible with Linux/QEMU UEFI VM installs, HDR 10bpp Nexus 6 test imagery, and Flask REST services for remote capture and display, shrinking manual validation loops
Built outdoor vision datasets (~12,000 images) with capture/annotation, Python RAW-color tooling, and rig control for multi-viewpoint experiments
Sustained imaging lab throughput with QNAP-backed Ubuntu workstations, networking support, and Linux/Windows servers powering experiments and REST tooling
Technical Skills: Python, C++, CUDA, OpenCV, Flask, Linux
Soft Skills: Research discipline, clear lab documentation
Describe architecture in a chat interface, uml-mcp MCP tools return a UML diagram
uml-mcp is an open-source MCP server for generating UML diagrams from chat. You describe the architecture, and the tool returns diagrams that are useful for reviews and documentation, with Mermaid, PlantUML, or Kroki. The same workflow runs in Cursor and in ChatGPT via a Custom GPT.
A docker compose up -d stack for standing up a media library on a Linux homelab, with downloads, *Arr automation, Plex or Jellyfin, remote access, and monitoring. The companion bash scripts help bootstrap a fresh host one service at a time before moving to the full compose setup.
Firmware for an ESP32 FireBeetle paired with an Adafruit I2C 7-segment display. A small local web UI lets you send digits and characters from any browser on the LAN.
Designed and published AR lenses on Snapchat (42 lenses, 2017–2020), with a focus on visual effects, fast iteration, and mobile reach. All-time Lens Insights: 13.97M plays, 20.38M views, 711.8k shares, 10,092 favorites, 753 posts, 428 unique posters.
ETSim contribution · packaging, release workflow, and data tooling
Accepted work on the ETSim codebase around package structure, release flow, and dependency submission. The project sits closer to master's-level research tooling than to club infrastructure. Also contributed TrueMap support to fredericjs/surfalize for surface-metrology workflows.
Built the official Annual Quebec-Ontario Pre-SIGGRAPH Workshop website, with a straightforward experience for the program, logistics, and attendee questions.
Contribution to anisotropic friction work in the IPC Toolkit. The work extends contact friction handling in a serious simulation codebase, across the C++/Eigen core, Python ipctk bindings, notebooks, and tests.
C++EigenPythonSimulationFriction
Club and school community work
Accepted contributions tied to student community projects and internal engineering initiatives.
AlgoÉTS contribution · AI financial-news backtesting
Contribution to an AlgoÉTS open-source FastAPI and MongoDB stack that ingests more than 60 sources, classifies news by sentiment and ticker, and tests how headlines relate to market moves. The project uses a Poetry/Docker workflow and public documentation.
AlgoÉTS contribution · NLP movie similarity with pgvector
Gradio app for title, multi-title, and rating-based queriest-SNE view of embedding neighborhoods2D projection of the movie embedding space
Open-source AlgoÉTS coursework on movie similarity: Sentence Transformers embeddings in PostgreSQL with pgvector, plus optional Qdrant and Neo4j tracks. Jupyter notebooks compare cosine and L2 distance, and a Gradio UI runs similarity search from titles or ratings.
PythonPostgreSQLpgvectorNLPGradioAlgoÉTS
Maintained community integrations
Forks and accepted contributions that remain useful to a broader open-source user base.
Maintains a .NET sink that connects Serilog to Sentry, so logs and errors can be centralized without heavy setup. Six package ids on NuGet cover net6.0 and net10.0 (main, ASP.NET Core, and framework-specific lines), with CI publishing on GitHub Releases and ready-to-use configuration examples.
Home Assistant (HACS) integration for Renpho smart scales
Vitals, weight history, and body-composition gaugesGirth measurements and goals from the integration
Forked and now maintain an open-source integration from neilzilla/hass-renpho that brings Renpho weight, BMI, body fat, girth measurements, and goals into Home Assistant. I reworked the API client, the config flow, optional proxy support, and the related documentation.
Home AssistantPythonHACSIoTOpen source
Other accepted contributions
A short list of merged work that stays off the main cards, but still shows range.
Have a role to fill or a project in mind? Available for software engineering roles and focused consulting mandates across platform, cloud, and graphics-heavy products. 24h response.
Recommendations. One recent quote is shown below, and full recommendations are available on LinkedIn.
“Antoine stands out not only for his technical expertise, but above all for his relational intelligence. He shows that an excellent developer can also be an exceptional teammate and a real boost to team morale.”
