Resume

Omar Chenine

AI Engineer and Data Scientist

Remote

Fourth year Data Science Engineering student. I build AI systems end to end: data preparation, modeling, retrieval, evaluation, and the interface. I care about baselines, leakage, and what happens when the model is wrong.

Download PDF Contact

Selected projects

RAG with retrieval evaluation

Python, retrieval, evaluation

Built a staged retrieval pipeline with a frozen evaluation set and cited answers.
Compared retrieval against a baseline before adding generation.
Tracked corpus size, chunking, top k, and reranker settings for each run.

ECG image analysis

Python, NumPy, signal processing

Digitized ECG images into a signal and extracted rhythm, axis, PR, and QRS.
Generated a clinical format report with a review flag for low confidence cases.
Documented calibration assumptions and medical limits.

Ames housing stacked ensemble

Python, scikit-learn, XGBoost, LightGBM, CatBoost

Built a stacked regression pipeline with log target, outlier policy, and ordinal encoding.
Fit preprocessing inside the fold and used out of fold predictions for the stack.
Compared the stack against single model and baseline scores.

BigQuery RFM and Power BI

SQL, BigQuery, Power BI

Wrote an RFM segmentation pipeline in BigQuery with explicit segment rules.
Built a Power BI dashboard for segment size and customer behavior.
Documented a warehouse and dashboard region conflict and the resolution.

Voice agents and AI agent work

Python, agents, voice

Designed tool using agents with typed inputs, fallback paths, and trace logs.
Set a per stage latency budget for voice turns.
Defined a task set and failure taxonomy for evaluation.

Sports outcome predictors

Python, pandas, scikit-learn, REST APIs

Built World Cup, F1, and boxing predictors with time based evaluation.
Used Elo as a baseline and cached API snapshots.
Reported calibration and sample size limits.