Machine Learning & Backend Developer

I design and deploy data-driven backends with a focus on recommender systems, retrieval quality, and low-latency serving. My background in business + IT (HEC Montréal, bilingual program) helps me align technical work with product goals and measurable impact.

Recently, I built a production book recommender end-to-end: normalized SQL schema, collaborative filtering for warm users, attention-pooled subject embeddings for cold start, and a FastAPI service with daily retraining and zero-downtime hot reloads. The system blends signals (ALS, embeddings, metadata) and prioritizes reliability and speed.

  • Python
  • FastAPI
  • PyTorch
  • LightGBM
  • FAISS
  • ALS
  • Scikit
  • SQL
  • Linux/Nginx
  • Azure
  • Cloud
  • Git
Projects Links About

Currently exploring: building on my recommender system experience by studying O’Reilly’s Data Engineering Design Patterns to strengthen my understanding of data engineering practices.

Recommenders

Hybrid pipeline: ALS (warm) + subject-embedding similarity (cold) with LightGBM reranking.

APIs & Serving

FastAPI endpoints with caching, pagination, and observability; deployed behind Nginx.

Data & MLOps

SQL modeling, feature preparation, daily retraining, artifact promotion, and safe hot reloads.

Featured