Houssem Ben Salem

I like math and I like software, so naturally I ended up in quantitative finance as a hobby. What started as curiosity turned into Lumin, a systematic signal model for crypto futures. It mines statistically robust price patterns from historical data, validates them out-of-sample, and deploys them across four timeframes (1d, 8h, 4h, 2h) on Binance perpetual futures. The current database holds 3,854 validated signatures. The 2026 out-of-sample results so far: +41.7% return, 84.5% win rate, 11.37 Sharpe, on four symbols starting from $500K simulated capital.

I built a terminal interface to explore the live model results.

I joined SYLVAC SA in Crissier, Switzerland, as an applied optics engineer building vision metrology systems — optics, illumination, cameras, image processing, motion control. In 2018 I moved to project management, owning end-to-end delivery of optical measurement machines: specs, industrialisation, production handoff. Since 2021 I lead the R&D team, making machines more intelligent and more autonomous — embedded agents, adaptive workflows, UX for robots. Machines should and will interact with humans better and more naturally.

Watch the testimonial →

Machine vision · optical metrology · quality control · robotic guidance · automation

Before SYLVAC, I was an applied optics engineer at Hexagon Manufacturing Intelligence. I worked on metrological machines, optical sensor development, and image processing for precision measurement systems.

Applied Data Science: Machine Learning program at the EPFL Extension School. This is where the quantitative finance rabbit hole started.

BSc + MSc MicroEngineering — Applied Optics at École Polytechnique Fédérale de Lausanne (EPFL). Five years of optics (vision), photonics (light guides and lasers), image processing and software.

I cross domains for fun. Medical devices, precision metrology, machine learning, quantitative finance — the pattern is the same: understand the physics, build the math, write the code, ship the thing. I’m an optimist by default, detail-obsessed by training, and I tend to stay on problems longer than most people think is reasonable, without losing efficiency.