At Mercor, I deliver batch and streaming data pipelines and LLM evaluation platforms using Python, SQL, Airflow, Spark, and Kafka. I developed event-processing pipelines with Bronze, Silver, and Gold data layers.
At École des ponts ParisTech, I built an end-to-end ETL pipeline processing more than 6.35 million Paris Vélib availability and weather records. I also developed ingestion workflows with Parquet, Google Cloud Storage, and BigQuery, and trained NLP models including spaCy NER and BERT.
Earlier, at Delante, I developed models for customer conversion and campaign performance, and at Adro Ad Exchange, I analyzed marketing data and developed dashboards to track advertising performance. My work spans data engineering, analytics, and machine learning, with projects including an LLM evaluation and monitoring platform and a luxury retail streaming platform.

