In August 2024 I started a second degree, a Master’s in Management for Engineers at the Higher School of Commercial Studies (EHEC) in Koléa, as part of a double-degree program with ENSIA.
Running Gostu and pitching at the Ministry had shown me that good engineering isn’t enough if you can’t explain why it matters, price it, or sell it. EHEC gave me the vocabulary for that.
Freelancing for Systelium
From June 2025 to June 2026 I worked remotely as a freelance AI & software engineer for Systelium, a company in France. I built a full-stack construction-management SaaS to the client’s requirements, including agentic AI features. Working to someone else’s specification, across a time zone and a language, was a different discipline from building my own products.
I also built CTC Flood Classification, a platform for engineers to collect and map nationwide project data for flood monitoring.
2025: the competitive year
2025 was the year I competed the most, and it went well:
- National AI Olympiads: first place, with DroughtFight, which combined time-series forecasting with AI-agent recommendations.
- S.H.E.S.E.M.: first place, with AgriScan, agricultural monitoring with on-device plant-disease classification.
- National Robotics and Programming Competition: second place and Best Software Implementation, also with AgriScan.
- Hack & Train: second place, across remote sensing, segmentation, and cybersecurity challenges.
- Mathurance Hackathon: third place, with an insurance provisioning prediction system.
The question that stuck
All of these projects used AI to help write, debug, or reason about code, and I kept wondering how reliable those tools really were. When an AI agent fixes a bug, how do you know the fix is right? How much does it cost? Is an open-source agent as good as a proprietary one?
I wanted to answer those questions properly, not just in a hackathon pitch. That’s what took me to Montréal.