ashref.devAI & Data Science Engineer / Blog

About
Projects
Now
Blog
Print One-Pager

The blog

Thoughts, projects
& things I learn.

Loading articles…

BERBAOUI Ashref / ENSIA · Algiers, Algeria

BUILT ON
PRECISION.

4th-year AI & Data Science Engineering Student at ENSIA. Open to impactful research partnerships and engineering opportunities.

Explore

AboutAAbboouutt ProjectsPPrroojjeeccttss NowNNooww BlogBBlloogg Print One-PagerPPrriinntt OOnnee--PPaaggeerr

The story behind the work

From Equations to Deployed Reality.

Every complex system starts with an equation, an observation, or an unmet need on campus. When I entered ENSIA, I realized that true engineering is not confined to Jupyter notebooks; it lives where mathematical models withstand real-world distributions, edge compute limits, and human usage.

Whether predicting the remaining orbital lifespan of a nano-satellite battery with ASAL, forecasting cafeteria meal demands to eliminate food waste for BNP Paribas, or building ENSIApedia to empower hundreds of peers, my focus remains constant: engineering with rigor, curiosity, and measurable integrity.

Elsewhere

GITHUB ↗GGIITTHHUUBB ↗↗LINKEDIN ↗LLIINNKKEEDDIINN ↗↗DISCORD ↗DDIISSCCOORRDD ↗↗EMAIL ↗EEMMAAIILL ↗↗

Get in touchGGeett iinn ttoouucchh ↗

© 2026 BERBAOUI Ashref. All rights reserved.AI Engineering & Applied Machine Learning.Back to top ↗BBaacckk ttoo ttoopp ↗↗

Technical Notes / BERBAOUI Ashref

THINK.
BUILD.
VERIFY.

✳

Reflections on time-series demand forecasting, deploying computer vision at the edge, and engineering reliable systems.

Loading articles…

3 published articlesNotes from the workbench

Time-Series Demand Forecasting with Gradient Boosting: Lessons from GDG AI Fest to BNP Paribas

Time-Series Demand Forecasting with Gradient Boosting: Lessons from GDG AI Fest to BNP Paribas

How feature engineering with cyclic calendars, lag matrices, and blended LightGBM + CatBoost models beat traditional baselines in commercial demand prediction.

BERBAOUI Ashref[views placeholder]

Deploying Computer Vision at the Edge: Real-Time Driver Fatigue Detection on Raspberry Pi

Deploying Computer Vision at the Edge: Real-Time Driver Fatigue Detection on Raspberry Pi

Balancing inference latency and precision: implementing eye-aspect-ratio (EAR) and head-pose estimators on resource-constrained embedded hardware with Qareeb.

BERBAOUI Ashref[views placeholder]

Architecting ENSIApedia: Engineering Knowledge Systems for Undergraduates

Architecting ENSIApedia: Engineering Knowledge Systems for Undergraduates

Why we built a dedicated digital knowledge repository for ENSIA engineering students, structuring course documentation, search indices, and collaborative study workflows.

BERBAOUI Ashref[views placeholder]