Hello, I'm
Building Intelligent Systems for Real-World Impact
Engineering Student @ CentraleSupélec & McGill University
AI & Machine Learning · Research · Multilingual NLP
I am a student in the Bachelor of Global Engineering at CentraleSupélec and McGill University, with a strong foundation in mathematics, data science, and software engineering.
I enjoy designing and building intelligent systems, turning ideas and data into practical tools. I love bringing ideas to life through projects and hackathons, where experimentation, iteration, and creativity meet real-world problem solving.
Through these experiences, I continuously explore how machine learning and AI systems can be used to build meaningful and impactful solutions.
What I'm working on right now
Running cross-lingual evaluations of frontier LLMs on olympiad-level linguistics benchmarks (MIST, IOLBench). Probing test-time scaling strategies and reasoning patterns under multilingual conditions at MILA, supervised by Julia Kreutzer (Cohere Labs).
Multi-agent AI system that analyzes airline and logistics cases to determine compensation eligibility and auto-generates claim submissions using LLM-based reasoning.
View on GitHub →Real problems. Real systems. Real impact.
Autonomous multi-agent system that monitors Gmail, detects compensation cases (flight delays, lost luggage, damaged parcels), evaluates eligibility under EU regulations, and automatically drafts and submits claims. Built with human-in-the-loop validation, browser automation (Playwright), and automated billing (Stripe). Stack: LLM reasoning (Claude), FastAPI backend, React dashboard, Supabase, Gmail API.
Modeled geospatial and operational constraints to identify optimal radar locations. Applied algorithmic filtering on terrain and distance constraints using elevation data. Generated coverage and line-of-sight analyses to support engineering decisions in collaboration with Thales.
Developed OOP control software for a Raspberry Pi–based mobile robot, implementing real-time line-following and trajectory tracking using sensor feedback and PID-based motion control. Integrated sensor inputs and motor drivers within a modular architecture to enable stable navigation and obstacle avoidance. Iteratively tuned control parameters to improve tracking accuracy and motion stability.
ML pipeline predicting train dwell times for improved public transport reliability in India. End-to-end: EDA, feature engineering, model fine-tuning.
Built an AI-powered marketing intelligence platform that aggregates social media trends, competitor activity, and customer feedback into a unified interactive dashboard. The system identifies high-impact channels, generates content suggestions, and supports data-driven budget decisions. Developed during Infosys BizHacks (24h) using multi-agent workflows in n8n, external APIs, and a custom chatbot interface.
Designed a federated learning–based AI surveillance framework for retail environments, enabling real-time detection of suspicious behavior without transferring raw video data. Implemented a distributed training architecture to preserve customer privacy while continuously improving detection performance across stores. Developed as a scalable solution for large retail chains such as Les Mousquetaires, combining computer vision, secure aggregation, and privacy-aware model updates.
Developed a numerical model in Python to simulate the Atlantic Meridional Overturning Circulation (AMOC), a key ocean current system influencing global climate stability. Built using NumPy, SciPy, and Matplotlib as part of CentraleSupélec's Modelling Week, applying computational methods to explore climate dynamics and tipping-point behavior.
Where I've worked and what I've built
MILA – Quebec AI Institute / Cohere Labs · Montreal, Canada (Remote)
Supervised by Julia Kreutzer, Senior Research Scientist
Infosys · Bangalore, India
Onepoint · Nantes, France
My academic journey
Bachelor of Global Engineering
Paris, France · Montreal, Canada
General Baccalaureate — Highest Honors
Marrakesh, Morocco
A snapshot of my technical and human toolkit
Rita's AI Assistant
MEET MY ASSISTANT
I built a Claude-powered assistant trained on my background. Ask it about my projects, what I'm looking for, or anything else.
I'm open to internships, research opportunities, hackathon collaborations, and interesting projects. Reach out, I reply fast.
GitHub
rita-berradaLocation
Paris, France · Montreal, Canada