Generative AI & multi-agent systems
LLM orchestration, RAG, embeddings, semantic comparison, and human-in-the-loop workflows.
What I do
Areas of expertise
Financial risk analytics
Applied machine learning and document intelligence for traceable, decision-ready risk analysis.
Data platforms & MLOps
Reproducible ETL pipelines, model lifecycle tooling, APIs, containers, testing, and CI/CD.
Responsible & explainable AI
Reliable systems designed for transparency, validation, and communication with non-technical leaders.
Professional journey
Experience
Data Scientist, AI & Risk Management · Intern
Desjardins · Risk Management Analytics
Designed Vigie Bancaire, an AI solution that automates quarterly comparative analysis of financial disclosures from Canada’s six largest banks. Architected a multi-agent table-matching pipeline and a semantic comparison engine, with human validation for traceability and reliability.
Impact
Reduced an estimated 5–10 days of manual analysis to a few hours per cycle.
Data Science Trainer
Orange Digital Center · Guinea
Led AI training for Orange Guinea employees on machine learning, natural language processing, and transformer models.
Lead Organizer · Deep Learning IndabaX Guinea
Deep Learning Indaba
Lead end-to-end planning for Guinea’s AI conference, coordinating sponsors, partners, speakers, mentors, and a hands-on machine learning hackathon.
Selected work
Projects
Data engineering
OpenFoodFacts Data Platform & Advanced Search
Built an end-to-end platform ingesting the full OpenFoodFacts dataset through scheduled Airflow ETL. Designed a multi-criteria search engine combining fuzzy matching and multifactor relevance scoring, exposed through FastAPI and a responsive Dash application.
Airflow · MinIO · PostgreSQL · MongoDB · FastAPI · Dash · Docker
View project
Generative AI
Multi-Agent AI Software Engineering Assistant
Led a four-agent system that automates requirements gathering, architecture design, multilingual code generation, and deployment configuration. Supports Gemini, OpenAI, Groq, and Ollama, with a Streamlit interface and 33-test pytest suite.
LangGraph · LangChain · LLMs · Streamlit · Docker · GitHub Actions
View project
Machine learning
Credit Card Fraud Detection
Ranked in the top 16% (100th of 641) in a Kaggle binary classification challenge using LightGBM on imbalanced tabular data.
Python · LightGBM · Scikit-learn · Feature engineering
View Kaggle profileTechnical toolkit
Skills
Generative AI & NLP
LangGraph, LangChain, OpenAI, Gemini, Groq, Ollama, Transformers, Embeddings, Multi-Agent Systems, RAG
Data Science
Python, SQL, Scikit-learn, LightGBM, PyTorch, TensorFlow
Data & MLOps
Airflow, MLflow, Docker, GitHub Actions, FastAPI, PostgreSQL, MongoDB, Spark
Visualization & Apps
Dash, Streamlit, Tableau
Cloud & Environments
GCP, AWS, Linux, Bash
Academic background
Education
M.Sc. Computer Science
Université du Québec à Montréal (UQAM)
2024 – Aug 2026 · Data Science & MLOps
B.Tech. Information Technology
Sharda University
2018 – 2022
Highlights
Recognition
- 2nd placeData Science Week Guinea 2022
- Country winnerUmojahack 2022 & 2023
- CIFAR Inclusive AI ScholarCanadian AI 2025
- Kaggle ExpertOpen data contributor
Research
Publication
An Efficient Approach for Predicting Financial Inclusion
IEEE ICCSEA 2022 · Machine learning for imbalanced financial inclusion data
Let’s connect
Interested in applied AI, data science, or MLOps?
I’m always open to discussing impactful projects and opportunities.