Paris, France
Internship Explorer
Internship Explorer
The Platform teams keep one of the largest computing platforms in the AdTech world functioning like clockwork. They keep our products running using a broad selection of technologies, like large scale data compute & storage services (Hadoop, SQL & NoSQL), streaming (Kafka), platform as a service (Chef, Mesos), identity management (Kerberos) and analytics (Hive, Druid, Vertica), as well as an extensive monitoring/observability infrastructure.For the Internship, you will be in a team of 5-7, working closely with your mentor to drive your project, design and ensure best practices are applied. You can ask questions and participate in all knowledge sharing sessions/workshops, etc. You are encouraged to actively voice your ideas whilst learning how to build and ship quality code into production which will likely affect millions of users instantly. During your internship (6 months) and according to your choice, skills and interest, you can tackle one of the following subjects/ teams:Observability: Select, test and integrate a reporting tool with the current stack: Prometheus / Graphite / Grafana / Elasticsearch / Kibana. Migrate Grafana to containers and integrate with SSO. Build a log streaming interfaceData Processing: Be part of a team that builds our Bigdata Flow platform and writes code to provide insight, give the platform users info about changes impacting their datasets (2) and even hint them about optimization opportunities.Distributed System SDKs: Smart cache invalidation in a distributed system.Continuous Deployment: Implement a mutation testing solution that is integrated into the Criteo CI/CD pipeline.Product Reliability Engineering: Migrate admin handlers' UI to Angular and help develop a load testing pipeline.Rivers: Create a Streaming Portal UI.Data Development Cycle: Leverage the data that we scrape from all our data processing systems to provide automatic monitoring and alerting and in-depth analysis to data producers so that they can understand the sources of delays and make better decisions on the design of their pipeline dependencies.
