Sacha Lasry

Contribution to the development of Trubrics, a leading product analytics platform for AI applications: ● Development and improvement of FastAPI endpoints to manage data ingestion and automated tagging process ● Integration of a front-end interface (TS, React) allowing users to easily map ingested traces with Trubrics expected format ● Design and implementation a topic modeling system using LLMs to extract insights from user-AI conversations ● Improvement of codebase structure, implementing CI/CD workflows, and enhancing development practices ● Introduction to AI code generation best practices to boost delivery speed and team productivity ● Integration of third-party APIs (Stripe, Langfuse, ...) to expand platform capabilities, allow seamless data synchronization and analytics across multiple services while maintaining data integrity, and bring new users to the platform

Support of Machine Learning team to improve their ML pipelines and development practices: ● Design and implementation of ML team dedicated repository and AWS infrastructure, while improving code quality and security ● Acceleration of data scientists development velocity through tool implementation (Sagemaker, ZenML, MLFlow, Cursor, uv ...) ● Optimization of the deployed LTV-prediction pipeline, reducing daily compute time from 5 to 2 hours

Analyses and development of dedicated solutions for efficiency improvement within company’s warehouses: ● Deployment of an MLFlow server allowing data scientists to track the results of their experiments on machine learning models ● Optimization of DBT models to reduce data warehouse refresh costs: Division by 3 of query costs ($36k yearly saved)