Omar Baker

Hey, I'm Omar

AI/ML Engineer Data Analyst Researcher

I build systems that learn — from self-driving cars to medical imaging and LLM fine-tuning. Based in Nablus, Palestine. 3+ years of applied ML experience, 15+ shipped projects, and 3 research papers in progress.

Get in touch GitHub LinkedIn

Experience

2025 — Present
Data Analyst
MT · Hybrid
Interactive Power BI dashboards, advanced DAX formulas, Power Query transformations. Turning complex business data into actionable KPIs for cross-functional teams.
2024 — 2025
Data Analyst
Unify · Remote
Predictive analytics models deployed via Python APIs. End-to-end data pipelines — collection, cleaning, visualization — integrated into client-facing platforms.
2022 — 2024
ML Engineer & Data Analyst
SHAI for AI · Nablus
Feature engineering pipelines on large-volume datasets. Stakeholder-facing data visualizations and detailed summary reports in a professional ML environment.

Projects

1st Place — Jordan & Egypt

Diamond Price Prediction

Won a regional ML competition for the best predictive model. Advanced regression, feature engineering, and model optimization for accurate market price forecasting.

Scikit-learn · Pandas · Regression · Feature Engineering
Automatic Essay Grading
Fine-tuned Mistral-7B with BERT as frozen encoder. Chain-of-thought prompting. MAE: 0.41/4 (10.34%).
Mistral-7B · BERT · Fine-tuning
Self-Driving Car
Behavioral cloning CNN in Udacity env. Image augmentation pipelines for steering angle prediction.
TensorFlow · CNN · OpenCV
CT Scan Segmentation
U-Net for semantic segmentation of medical CT images with custom preprocessing pipeline.
U-Net · TensorFlow · OpenCV
Fraud Detection
ML-based financial fraud detection with risk scoring for real-time transaction classification.
Python · Feature Eng. · Analytics
Text Summarization
Custom transformer from scratch for abstractive summarization on news articles.
Transformers · Hugging Face · NLTK
Time-Series Forecasting
LSTM vs Prophet for stock and weather prediction with forecast accuracy analysis.
LSTM · Prophet · TensorFlow
Pothole Detection
DeepLabV3 for road damage segmentation with custom dataset annotation and augmentation.
DeepLabV3 · OpenCV · Segmentation

Research

[1]
Enhancing LLM Understanding through Embedding & Instruction Tuning
Integrating embeddings with instruction tuning to improve semantic comprehension and response accuracy in large language models.
In progress
[2]
Genetic Research on Lung Cancer Genes
Bioinformatics and ML analysis of BAX, BAD, GSTP1, BCL2 gene expression for biomarker identification.
In progress
[3]
Fraud Detection System Design
Scalable AI/ML architecture for predicting and preventing fraudulent financial transactions.
In progress

Skills

ML & AI
TensorFlow Keras PyTorch Scikit-learn Hugging Face NLTK Prophet
Data
Python SQL Pandas NumPy Power BI DAX Plotly
Vision
OpenCV Dlib YOLOv5 U-Net DeepLabV3 OpenPose
Tools
FastAPI Git HTML/CSS Power Query Beautiful Soup

Education

BSc Artificial Intelligence
An-Najah National University — Nablus, Palestine
AI · Data Science · Machine Learning · Deep Learning · Neural Networks · Statistical Analysis
Advanced Python
Gaza Sky Geeks · 2024