Data Scientist & AI Engineer

HAMZA
CHIKHAOUI

Hi there! My name is Hamza Chikhaoui and my passion lies where engineering meets machine intelligence.
Currently building AI systems at Pratt & Whitney

✈ Aerospace AI 🧠 Machine Learning 🔗 Graph Neural Networks ⚙ GenAI / RAG McGill '25
Hamza Chikhaoui
Currently @ Pratt & Whitney DATA SCIENTIST· MONTRÉAL, QC
SCROLL
About Me

THE MECHIE WHO CODES

I'm a mechanical engineer turned AI scientist — a combination that lets me see both the physical systems and the data that describes them. Trained at McGill University - MILA in applied mathematics and machine learning, I care about pushing the boundaries of what the latest models can do — but always in service of solutions people will actually rely on.

My two biggest passions — aerospace and artificial intelligence — converged at Pratt & Whitney, where I build GenAI and ML applications that solve real operational problems and make engineers' work meaningfully easier.

I also co-founded Clementine.io, a startup using graph machine learning and GenAI to help small and medium businesses act on their data. Most companies aren't short on data — they're short on the tools to turn it into decisions. That's the gap we close.

I'm fluent in four languages, love exploring new cities with a camera in hand, and believe great engineering always starts with genuinely understanding what someone needs.

// Core Stack

PythonPyTorch LangChainLlamaIndex LangGraphDatabricks AWS BedrockPyTorch Geometric HuggingFaceSpark SQLDSPy LanceDBAzureDevOps
4.0
MSc GPA
51+
Software Implementation Projects Led
400h
Client Consulting Hours
4
Languages Spoken Fluently
McGill Degrees + James McGill Scholar
Career

WHERE I'VE WORKED

Building at the intersection of aerospace, enterprise AI, and startup innovation.

✈ Pratt & Whitney May 2025 — Present
Montréal, Canada
Data Science Intern
  • Engineering end-to-end RAG and GenAI tools to streamline operations across business units — from cloud ingestion through model development to production UI.
  • Designing scalable cloud architectures for AI solutions at one of the world's leading jet engine manufacturers.
  • Applying state-of-the-art language models and retrieval systems to real aerospace challenges every day.
🍊 Clementine.io Sep 2024 — Present
Montréal, Canada
Co-Founder & Lead Data Scientist
  • Co-founded a startup transforming relational database analysis with graph machine learning for personalized business intelligence.
  • Developed GNN-powered recommendation systems enabling data-driven decisions through advanced network analysis.
  • Full ownership from model architecture to product delivery.
🔧 MaintainX May 2023 — Sep 2023
Montréal, Canada
Software Implementation Consultant
  • Led 51 software implementation projects end-to-end, troubleshooting technical issues and optimizing system performance.
  • 400+ hours of consulting — data imports, normalization, visualization, and user training.
🎓 McGill University 2019 — 2025
Montréal, Canada
MSc Computer Science + BEng Mechanical
  • MSc specialization in Machine Learning & AI at McGill & MILA. GPA: 4.0 / 4.0.
  • BEng Mechanical Engineering with Minor in Applied AI. GPA 3.73/4.0. James McGill Scholarship recipient.
Portfolio

SELECTED PROJECTS

Machine learning projects spanning NLP, graph learning, computer vision, and interpretable AI.

🔬
BioNetRec
GNN-based recommendation system for medical specialists — analyzes biomarkers and patient history. Features an LLM-powered chatbot with guardrails for personalized, interpretable recommendations.
PyTorch GeometricGNNLLMHealthcare AI
💬
BERT Code Generator
CFG-based code generator translating natural language to executable commands with visual outputs. Sentence-BERT semantic embeddings achieving performance comparable to fine-tuned LLMs.
BERTNLPCFGCode Gen
🔩
Screw Defect Detection
6D Convolutional Neural Network in a few-shot learning environment for industrial screw defect detection using the ProtoNet meta-learner. Robust to low-data industrial regimes.
CNNFew-ShotProtoNetCV
📊
Explainable XGBoost
LIME and SHAP explainers to analyze global feature attributions for an XGBoost classifier — enhancing model interpretability and decision-making transparency.
XGBoostSHAPLIMEXAI
🗳️
LLM Political Persona Classifier using DSPY
Investigated whether DSPy's constrained generation can improve LLM classification of voting personas (Democrat vs. Republican) from policy preferences and tweets. DSPy optimization yielded a +32 percentage point accuracy gain over the baseline LLM — while also partially correcting inherent political bias.
DSPyLLMNLPClassificationBias Mitigation
🍊
Clementine.io Platform
Graph ML platform transforming relational database analysis into personalized business intelligence. GNN-powered recommendation engine for data-driven network decisions.
Graph MLGNNStartupB2B
View All Projects on GitHub →
Through My Lens

I AM PASSIONATE BY PHOTOGRAPHY

Here are some pictures I was lucky enough to be able to take — moments captured between engineering and exploration.

The Human Side

BEYOND THE CODE

I believe the best engineers are curious humans first. I grew up between cultures — with roots in North Africa, and wings in Montréal — and this has shaped how I approach problems: with empathy, adaptability, and a global perspective. I love exploring new cities, a camera in hand, and finding the stories that connect people across borders.

🇨🇦
Home Base
Montréal, Canada
Where it all began. The city that shaped my love for photography, jazz, and engineering.
🇲🇦
Roots
Morocco
Where I was born, and where my family is from. The medinas, the Atlas Mountains, the warmth of Moroccan hospitality.
🌱
Community
Refugee Center McGill
Founded & VP'd the chapter — mentoring a team of 50, supporting hundreds of recently-admitted refugees.
🌍
Exploration
Europe & Beyond
Exploring different engineering cultures and perspectives across cities, one trip at a time.
🇬🇧
EnglishFLUENT
🇫🇷
FrenchNATIVE
🇲🇦
ArabicNATIVE
🇪🇸
SpanishINTERMEDIATE
Get In Touch

LET'S CONNECT

Whether you want to discuss AI in aerospace, collaborate on a project, or just say hello — I'm always happy to connect.