Skip → résumé

Wael

ML engineer & AI-agent builder, competing on Kaggle & Zindi.

I build forecasting models and multi-agent LLM systems, and compete on Kaggle and Zindi. Open to collaborations on machine learning, AI agents and data competitions.

Résumé Contact

A small explorable underground: every room is a project. Arrows/WASD + Z. Sound is off.

Wael

ML engineer & AI-agent builder, competing on Kaggle & Zindi.

I build forecasting models and multi-agent LLM systems, and compete on Kaggle and Zindi. Open to collaborations on machine learning, AI agents and data competitions.

This portfolio is a small game and needs JavaScript. Everything is also on the résumé page.

TrafficFlowBench 2026: Causal traffic-state reconstruction & queue forecasting under a hidden regime shift.
The Refactoring Swarm: Multi-agent LLM system that detects, fixes and validates Python code issues in a self-healing loop.
One Road Out: Zindi Bias Bounty: an escape-route road filter hides second exits for 2.88M residents in wildfire country.
ML Algorithms From Scratch: Core machine-learning algorithms from first principles.
VEC Leaderboard Insights: Which metric should I improve next? A tool for the NeurIPS 2026 Virtual Embryo Challenge.
LLM-FineTune (Qwen 2.5 3B QLoRA): Fine-tuning Qwen 2.5 3B with QLoRA on a cleaned UltraChat corpus.
APCS Frontend (MicroHack): Hackathon frontend for booking container-terminal appointments, with admin, operator and carrier portals.