Data Science · AI / ML · Games

Elie Khoubbieh
Building systems
that think

I design intelligent systems at the intersection of machine learning, data science, and interactive experiences.

Scroll

About

Who I am

I'm a data scientist and ML engineer with a passion for building things that bridge analytical rigor and creative interactivity. Whether training models on complex datasets or prototyping game systems, I love problems at the edge of disciplines.

My work spans predictive modeling, reinforcement learning, and real-time data pipelines, with a side of game mechanics and procedural generation.

Years exp.

Projects

Domains

Core skills

Data Pipelines

88%

Agentic Systems

79%

Python / ML

93%

Deep Learning

86%

Game Dev

74%

Visualization

81%

Work

Selected projects

Reinforcement Learning

ML Agents for Fire Escape Logic

Trained a PPO agent to navigate layouts, achieving human-level performance. The environment was built in Unreal Engine with a custom Python bridge for training. The purpose was fire safety research

PyTorchStable-Baselines3Unreal EnginePython

→

Data Science · Live Systems

Real-time Bike Sharing Demand Forecasting

End-to-end ML pipeline predicting bike demand and detecting anomalies or defects in certain bikes by analysing rental patterns. Includes a FastAPI inference endpoint with sub-20ms latency.

XGBoostKafkaFastAPIPostgreSQL

→

Generative AI · Computer Vision

Procedural Terrain via Diffusion Models

Fine-tuned a latent diffusion model on satellite imagery to generate realistic, playable heightmaps for open-world games. Outputs feed into a custom WebGL pipeline that renders terrain in real time, with biome classification from a secondary CNN.

DiffusersWebGLGLSLUnity

→

NLP · LLM Fine-tuning

Fine tuning LLMs for Optimized Gameplay

Fine-tuned an LLM on interactive Player vs Player games to quickly find optimal strategies and understand game limitations.

oLLAMALangGraphMCPHuggingFace

→

Contact