MSc Artificial Intelligence @ Imperial College London
Causal ML · Multi-agent systems · Recommender systems
AI engineer focused on systems that affect public and economic outcomes. Background: Business Information & Decision Systems → marketing & growth at noon (UAE's #1 Q-commerce platform) → MSc AI at Imperial. Open to graduate AI Engineer / Applied Scientist roles in London for 2026.
Thesis — algorithmic-meritocracy
NeurIPS-style paper testing the platform claim that "the algorithm surfaces the best content". An agent-based simulation (1k creators × 10k users × 100 timesteps) decomposes engagement into earned vs. manufactured components using inverse propensity weighting and covariate adjustment.
Headline: only 7–18% of engagement variance is uniquely quality-driven; 63–73% is absorbed into a feedback-loop-manufactured shared component. User-tunable algorithm weights Pareto-dominate the platform default by +26.6% rank correlation, and just 5% adoption captures 68% of the achievable gain.
Other work is pinned below — RAG over Gmail, a multi-agent dispute-resolution demo, an alt-data quant factor model, and microscopy deep learning.
Python (PyTorch, scikit-learn, pandas, FastAPI) · TypeScript / React · Causal libs (DoWhy, EconML) · Three.js · LLM agent frameworks · LaTeX
Fastest channel: jouzoumohamad@gmail.com. Happy to chat about anything causal, multi-agent, or e-commerce ML.