Nicolas Landier

I lead a team composed of Data Engineers, Data Scientists and Analytics Engineers which is in charge of:

- collect and gather the data onto our Snowflake Data Warehouse

- design and implement a data model to foster self-service analytics using DBT and Looker

- develop Machine Learning and automated approaches to improve the product leveraging the data

At Criteo, I joined the Data Processing team which was in charge of implementing and maintaining the data pipelines. I then had the opportunity to bootstrap a new team dedicated to designing & building the analytical Data Model to be used by all the company analysts (+300 data analysts world wide) and deployed on Vertica. Moving on I got tasked to lead a team dedicated to all the Data Exploration use cases: provide all the Data Analysts, Researchers and other Data Engineers with tools to enable them with new Machine Learning approaches. For instance, a project we handled was to build a self-service work environment on the internal cloud infrastructure - a la AWS EC2 - in order to maximize the GPU resources usage. Finally I assumed the lead for a team which was building an internal reporting portal for our customers so they could access their own data. For this, we implementend a Looker-like application that would give the autonomy to the customers in terms of reporting possibilities.

At Deezer, I was in charge of Data Engineering and Data Infrastructure which consists in daily data workflows execution and the data processing on a Hadoop cluster and some databases around. Some projects I have worked on with the assistance of a 5-people team I managed:

- rework of the whole Big Data project in order to decentralize the release process among the different data teams (royalties, recommendation, catalog, ...);

- introduction of Agile rituals and tools: stand up meeting, retrospective, workshops...

- animation of Community of Practices for Data developers and users;

- automation of Hadoop cluster deployment with Ansible;

- Hadoop migration from Cloudera to HortonWorks distribution;

- refactored central Data Warehouse tables around Spark in order to increase maintainability and code quality and implement job optimization around data ingestion and data processing (switch to Parquet and Avro storage formats, Hive tuning, set up Tez).

- introduction of automation tools at Deezer now used by the infrastructure team;

- introduction of Prometheus and InfluxDB for monitoring purpose now used by the Infrastructure team;

- introduction of Cassandra set up to replace MySQL databases to store user centric data for the website.