Skip to content
Antoine Boucher

Options

Profiles

← All articles
Article

My journey in software engineering

5 min read

Listen to the audio story

From Flash games to platform work · 2:25 · AI-written script, synthetic voices

Transcript

Ava: Antoine didn't start in Kubernetes. He started in browser tabs, game loops, and a makerspace that smelled like hot PLA.

Andrew: That's a better origin story than most. How far back does it go?

Ava: Around 2010, in high school, building small browser games in Flash and learning HTML and JavaScript. The habit was to ship something playable, then fix what feels wrong. Years later he packaged some of that nostalgia as a Flash games site and a Docker setup for classic consoles.

Andrew: Then bigger tools?

Ava: Around 2015, a portfolio of game projects redone in Unity, CryEngine and Unreal. It taught him scope. By 2017 he was managing a makerspace at a college, with 3D printers and VR gear, and he built a WiFi-controlled RC car. That's also when he got serious about web development.

Andrew: Where does the career part start?

Ava: 2018, a five-week internship at Algolux, in computational imaging. Python tools, RAW color analysis, capture workflows. It trained a loop he still uses: measure, automate, document.

Andrew: And school?

Ava: At ÉTS, a capstone hovercraft, a hydroglisseur, with a lift fan and 3D-printed parts. A bachelor's from 2018 to 2023, helping run the AlgoÉTS algorithmic trading club. Co-ops from 2019 to 2022 at Wandrian, Power Go and Intact, then full time in 2023 at IONODES on an IoT platform. And a master's from 2023 to 2026 on real-time rendering and surface wear.

Andrew: What changed after graduating?

Ava: The problems moved up a layer. GitLab CI/CD, GCP, Java microservices, and DevSecOps labs at Polytechnique Montréal, with GitLab templates for about twenty teams. Plus teaching at ÉTS.

Andrew: What stayed constant?

Ava: He used to count features shipped. Now he cares whether a system stays understandable when he's not in the room. Reliability is trust, and security is habits in the pipeline, not a final gate.

Andrew: Any advice?

Ava: Follow problems where you see the whole loop: code, deploy, operate, improve. His started with Flash and a hovercraft. The full write-up has the timeline diagram and project links.

My journey in software engineering
Table of contents

I did not start in Kubernetes. I started in browser tabs, game loops, and a makerspace that smelled like hot PLA. This post is the longer version of that path—backend, platform, and DevSecOps are where I landed, but the steps in between matter.

1. The arc (one glance)

Career flow (Mermaid source):

Career journey from Flash games and college makerspace to backend platform work today

Figure: milestones from browser games and makerspace demos through ÉTS capstones, co-ops, and today’s platform/backend focus. Rendered from career-journey-flow.mmd.

Color key: green personal · blue school · purple club · gray co-op · slate full-time job · orange contract · pink teaching · amber today.

The diagram is the map; the sections below are the stories I tell in interviews when someone asks “how did you end up doing platform work?”

2. ~2010 — Game developer energy, Flash and the web

In high school I was already building small browser games in Flash and picking up HTML and JavaScript. Progress meant something you could click: a loop that runs, a score that updates, a page that loads without embarrassing you in front of friends.

That habit—ship something playable, then fix what feels wrong—never really left. Years later I packaged some of that nostalgia as FlashGames[1] (Ruffle in the browser) and a https://antoineboucher.info/CV/blog/projects/retroarch-web-games/ Docker setup for classic consoles.

3. ~2015 — A game portfolio, then bigger engines

By college I was curating a portfolio of game projects and redoing favorites in Unity, CryEngine, and Unreal. Same motivation, heavier tools: lighting, physics, assets that take a week to look “fine.”

It taught me scope. A jam game and a semester demo do not owe each other the same bar—but both owe you honest playtesting.

That scope sense shows up later when I refuse to boil the ocean on a single microservice migration — I want a playable milestone every week.

4. ~2017 — Makerspace, WiFi RC, web as a serious path

At Collégial International Sainte-Anne I managed the makerspace: 3D printers, VR gear, students stuck on wiring at 4 p.m. on a Friday. I also built a WiFi-controlled RC car—the kind of project where firmware, mechanical slack, and “why does it only steer left?” share the same afternoon.

That year I leaned into web development for real: less “scene in a game engine,” more “product someone uses every week.”

5. 2018 — Algolux internship (computational imaging)

My Algolux internship was five weeks inside computational photography: Python tools, RAW color analysis, capture workflows, Ansible stacks for ISP validation. I still have the research poster from that period—it is the bridge between “I like code” and “I like systems that touch the physical world.”

Research poster — AI for image quality optimization, Algolux internship, Collégial Sainte-Anne

Concrete leftovers on this site: https://antoineboucher.info/CV/blog/projects/rawanalyser/ (RAW clipping and calibration helpers, Algolux-era lineage). Lab work was spreadsheets, cameras, and scripts—not shiny product pages—but it trained the same loop I use today: measure, automate, document.

6. ÉTS — Hydroglisseur, AlgoÉTS, and the web stack

École de technologie supérieure (ÉTS) is where school projects stacked into something like a career.

Technological path (2017–2018) — capstone hydroglisseur (hovercraft): blue skirt, lift fan, 3D-printed parts, wood and foam and stubborn debugging. The kind of build that teaches integration before microservices.

ÉTS hydroglisseur prototype — lift fan, wooden deck, and blue inflatable skirt on the workshop floor

Bachelor (2018–2023) — coursework plus clubs. I helped run AlgoÉTS[2] (algorithmic trading); we ran workshops like Python 101 for data science for members who wanted to move from “I use Python” to “I trust my pandas.”

AlgoÉTS Python 101 — Science des données session on a classroom display

That club work connects to open source I still maintain: the https://antoineboucher.info/CV/blog/projects/marketwatch/ Python library for the stock-market game. Same itch—automate the boring path so humans can focus on decisions.

Co-ops (2019–2022) — Wandrian[3], Power Go[4], Intact[5]: parsers, APIs, ELK dashboards, and backend work where the titles finally matched the daily grind.

Full-time (2023) — IONODES[6]: cloud developer on an IoT platform—Auth0 tiers, Sentry, ONVIF/WebRTC, Azure DevOps.

Master (2023–2026) — ÉTS: real-time rendering, interactive physics, and surface wear in real time, with dynamic friction and textures—the thread that ties graphics back to the path.

7. 2020s — Platform, security, teaching

After graduation the problems moved up a layer:

  • GitLab CI/CD, GCP, Java microservices, Playwright tests (IMC2 research).
  • DevSecOps labs at Polytechnique Montréal—GitLab templates for ~20 teams, Docker/Kubernetes, OWASP-minded reviews.
  • Teaching at ÉTS: distributed databases, mobile UX labs, integrator projects—courseware is also software delivery.
  • The GraphQuon conference website[7].

8. What stayed constant (the “human” part)

Early on I measured progress in features shipped. Now I care whether the system stays understandable when you are not in the room:

  • Backend — boundaries, failure modes, APIs that do not surprise the next team.
  • Platform — CI, logging, local dev that works without a heroic README.
  • DevSecOps — secrets, dependencies, and deploy paths that are boring on purpose.

Reliability is trust. Security is not a final gate—it is habits in the pipeline. I still write posts and draw flows because explaining something badly is how I notice what I do not understand.

9. What I optimize for next

Sharper platforms, safer delivery, legible systems as they grow. If you are early in the path: follow problems where you see the whole loop—code, deploy, operate, improve. I happened to start that loop with Flash and a hovercraft; you might start somewhere else. The loop is the job.

More concrete write-ups live under posts and projects on this site.

References

  1. FlashGames. (). [Web page] antoinebou12.github.io.
  2. AlgoÉTS. (). [Web page] algoets.etsmtl.ca.
  3. Wandrian. (). [Web page] wandrian.com.
  4. Power Go. (). [Web page] www.powergo.ca.
  5. Intact. (). [Web page] www.intact.ca.
  6. IONODES. (). [Web page] ionodes.com.
  7. GraphQuon conference website. (). [Web page] graphquon.github.io.

References use this article’s publication date.