Mohamed Ali Habib

Mohamed Ali Habib

AI/ML Engineer | Data Scientist

I build end-to-end AI systems for enterprise clients. From the sensor to the screen, I design IoT hardware, implement computer vision and GenAI/LLM models, and write the software that ties it together. Currently shipping: a CV system analyzing thousands of hours of video, scaling to 100+ sites in a major US city.

View My WorkDownload Resume

PROVEN IMPACT AT A GLANCE

1000s+
Hours of video processed in production
100+ sites
Deployment scale
End-to-end
Hardware → AI → dashboard
Hours → Mins
Computation time reduced
FareLens: CV/AI Fare Evasion Tracking

FareLens: CV/AI Fare Evasion Tracking

Computer VisionDeep LearningPython
Multi-modal GenAI + IoT: Bus Stop Asset Monitoring

Multi-modal GenAI + IoT: Bus Stop Asset Monitoring

Google GeminiCustom IoT deviceGoogle Cloud
Trip Scheduling Optimization

Trip Scheduling Optimization

PythonRubyOptimization
Interactive Anchor Tracking Tool

Interactive Anchor Tracking Tool

PythonVisualizationData Analysis
SCROLL TO EXPLORE

Experience

Hatch logo

Data Scientist

Hatch
March 2024 - Present
White River Junction, VT, USA
  • Built FareLens, a computer vision system for fare evasion tracking that analyzed 1500+ hours of video across 10 sites, scaling to 100+ sites in a major US transit network.
  • Created an end-to-end IoT + GenAI asset monitoring system: designed hardware to capture imagery and integrated multimodal Gemini LLM for automated analysis, in collaboration with Google Cloud team.
  • Rewrote trip scheduling optimization using constraint programming and graph algorithms, reducing computation time from hours to minutes and expanding features.
  • Built interactive data visualization tools for equipment installation tracking and stakeholder reporting.
  • Automated ETL pipelines for multi-source data integration, eliminating weeks of manual processing.
Computer VisionLLMIoTDeep LearningGoogle Cloud GCPOptimizationETLPython
University of Oklahoma logo

Graduate Research Assistant

University of Oklahoma
Aug 2021 - Dec 2023
Tulsa, OK, USA
  • Built ML models for road surface classification (0.9+ F1) and highway travel time estimation (90%+ accuracy) using 27 months of RWIS weather and radar speed data across 200+ miles on I-35.
  • Developed a real-time Android app deployed on 300+ Oklahoma snowplow truck fleet for driver communication and sensor data collection.
  • Developed statistical efficiency metrics to quantify snowplow clearance impact on traffic speed recovery across 64 storm trips
  • Programmed BeagleBone embedded system to collect real-time road surface temperature from truck-mounted IR sensors.
Machine LearningData ScienceIoTPythonTensorflowPyTorchAndroidBeagleBone

Software Engineer

Salvimar

Jun 2020 - Jul 2021
Remote work, Germany
  • Developed web applications (React, TypeScript, Node.js) and Android apps (Kotlin, Java) in an agile team environment.
  • Led agile projects ensuring code quality and team coordination.
React.jsTypeScriptNode.jsKotlinJavaMongoDBGraphQLAgile
Wizme Ltd logo

Data Engineer

Wizme Ltd
Jul 2020 - Oct 2020
Remote work, London, England, UK
  • Built Python/Scrapy web scrapers for data collection from diverse sources, improving ETL processes and data quality.
PythonScrapyWeb ScrapingData EngineeringETL

Projects

Enterprise AI solutions, IoT systems, and machine learning projects.

FareLens: CV/AI Fare Evasion Tracking

FareLens: CV/AI Fare Evasion Tracking

I built FareLens, a computer vision system that analyzes video footage to detect fare evasion incidents and measure ridership for transit agencies. The system analyzed 1500+ hours of CCTV footage across 10 stations in a large transit hub in the US and is now scaling to 100+ sites citywide.

1500+ hours analyzed
100+ sites
Computer VisionDeep LearningPythonDashboard Analytics
Multi-modal GenAI + IoT: Bus Stop Asset Monitoring

Multi-modal GenAI + IoT: Bus Stop Asset Monitoring

I built an end-to-end real-time system that enables transit agencies to monitor bus stop assets and evaluate conditions daily instead of 2-3 times per year. The solution combines a custom IoT device, camera, and cloud-based Google Gemini for multimodal visual analysis. Project in collaboration with Google.

Real-Time
365x frequency increase
Google collaboration
End-to-end
Google GeminiCustom IoT deviceGoogle CloudPython
Trip Scheduling Optimization

Trip Scheduling Optimization

A scheduling optimization tool that assigns train trips to station tracks for maximum capacity utilization. Built using constraint programming and graph-based optimization. Reduced computation time from hours to minutes while adding new analysis features. Used to analyze operational plans in one of the busiest transit hubs in the US.

Hours → Minutes
Max capacity optimization
PythonRubyOptimizationGraph TheoryAlgorithmsGoogle OR-Tools
Interactive Anchor Tracking Tool

Interactive Anchor Tracking Tool

A data visualization tool that allows tracking of equipment installation status to help monitor current progress and plan upcoming installations. Featured in MBTA (Massachusetts Bay Transportation Authority) board meeting presentations, enabling stakeholders to make data-driven decisions.

Featured in MBTA board meeting
Team collaboration tool
PythonVisualizationData AnalysisETL
MATT: Real-time App for Snowplow Trucks

MATT: Real-time App for Snowplow Trucks

A real-time IoT Android application deployed in Oklahoma state for ODOT 300+ snowplow trucks enabling seamless communication between drivers and dispatch. The app collects data from embedded systems and provides real-time updates on road conditions, helping improve winter road maintenance efficiency.

AndroidIoTReal-time
Travel Time Estimation Using Radar Sensor Data

Travel Time Estimation Using Radar Sensor Data

Developed machine learning models to estimate travel time on highways using radar and GPS-based speed data. Achieved over 90% accuracy through feature engineering and model optimization.

MLIoTPython+2
Road Surface Conditions Classification

Road Surface Conditions Classification

Implemented ML models to classify road surface conditions using RWIS weather and speed data, achieving a 0.9+ F1 score. Part of research analyzing 27 months of data across 200+ miles of highway.

MLPythonClassification+2
Brain Tumor Detection

Brain Tumor Detection

Developed a convolutional neural network to classify brain tumors into malignant and benign categories using medical imaging data. The CNN architecture built from scratch achieves 0.8 F1

Deep LearningComputer VisionCNN+1
Sarcastic Chatbot

Sarcastic Chatbot

A conversational AI chatbot with a sense of humor built using encoder-decoder LSTM architecture with Bahdanau's attention mechanism. The model was trained on dialogue datasets to generate contextually relevant and humorous responses.

Deep LearningNLPLSTM+2
Article Recommender System

Article Recommender System

1st place hackathon winner. Built a recommender system for Arabic articles using Doc2Vec and Word2Vec models. The system analyzes article content and user preferences to provide personalized recommendations.

MLNLPPython+2
Technical Expertise

Skills

Specialized in AI systems that process real-world data—from computer vision systems to LLM-powered applications solving enterprise challenges.

AI & Machine Learning

Building intelligent systems that see, learn, and decide

1500s+ hours of video processed
  • Computer Vision
  • LLMs/GenAI (Gemini)
  • Deep Learning
  • Python
  • PyTorch
  • TensorFlow
  • OpenCV
  • scikit-learn

Software Engineering

Production systems from algorithms to optimization

Optimization: hours → minutes
  • System Design
  • Full Stack Development
  • Algorithms
  • Optimization
  • Python
  • JavaScript/TypeScript
  • Docker
  • Google Cloud (GCP)

IoT & Hardware

End-to-end system development & edge computing

GenAI+IoT system on Google Cloud
  • Real-time Systems
  • Raspberry Pi
  • LTE/GPS Modules
  • Camera Systems
  • Sensor Integration
  • BeagleBone

Also experienced with

Google OR-ToolsCloud RunVertex AISQLNext.jsReactNode.jsTailwind CSSSaSSHTMLCSSRubyPostgreSQLMongoDBREST APIsGraphQLGoogle LookerPowerBIETL PipelinesData VisualizationGitKotlinJava

Education

University of Aleppo logo

Bachelor's Degree in Computer Science

University of Aleppo

2014 - 2019
Top 5% of class5-year degree program

Awards & Recognition

Oklahoma Department of Transportation (ODOT)
University of Oklahoma
Oct 2023

Oklahoma Transportation Research Day 2023

Oklahoma Department of Transportation (ODOT)

3rd Place Winner, OTRD Student Poster Competition in Oklahoma City. Poster: The Importance of RWIS Subsurface Temperature Probes.

ITS Heartland
University of Oklahoma
April 2022

ITS Heartland 2022

ITS Heartland

3rd Place Winner, Student Poster Competition Annual Meeting in Lincoln, Nebraska. Poster: Speed and Travel Time Estimation Analysis of Traffic Systems.

Wikilogia Hackathon
April 2018

Wikilogia Hackathon - Data Science and Machine Learning

Wikilogia Hackathon

1st Place Winner. Built easyLearn, an educational blogging website with Arabic text recommender using NLP and ML techniques.

KDnuggets
Sep 2019

KDnuggets Most Popular Post

KDnuggets

Awarded golden badge for the article: "Advice on building a machine learning career and reading research papers by Prof. Andrew Ng"

Presentations

Contact

|

Connect with me

Email
LinkedIn