Working on the multiply.cloud backend and core services.Main achievements:- AI powered translator tool- CI tool to parallelize tests across multiple containers- demo mode: e-commerce simulation for trial accounts- matching: customers can link their products with competitors products using suggestions- multi-server: instances with their own disjoint set of customers- third party marketplace api integrations: products and offers, price updates, sales, metrics, etc (amazon SP, google shopping, cdiscount, rakuten, manomano, mirakl...)- python3 migration- pricing algorithms for multiple marketplaces (cdiscount, manomano, backmarket…)- audit log: store diff and author of postgresql operations or API actions- blockly: users can build their own transformers for their catalog ingestion (sku, listing price, shipping price...) using a visual programming language- billing plans- catalog feeds with generic drivers for csv, xml, xlsx content- observations refactor: optimal products information fetching using scraping or external apis- notifications: end users emails and UI messagesTechnologies: python, javascript, postgresql, flask, fastapi, sqlalchemy, redis, docker

Worked on the data pipelines, from the external systems (mainframe, oracle, mysql, mongodb...) to the exposed tables, metrics and statistics.POC, technical specifications, development, packaging, deployment, internal documentation.Main achievements :- low latency data processing tool to replace obscure and mystic excel sheets (python, flask, pandas, mongo, gridfs, js)- spark CICD stack: spark test runner, packaging and deployment tools- product recommendation spark project- GDPR framework- email classification micro service (tech lead, design and core/backend dev)- base ETL spark job- ETL job creator: infer raw inputs and create the spark jobs- dynamic integration tests for ETL spark jobs- ETL jobs for various sources- raw input ETL files watcher- hadoop/spark probesTechnologies : python, spark, hadoop, oozie, mongodb, hive, impala, flask, apache httpd, jenkins, microsoft vsts

Worked on the data pipelines, from the events producers to the exposed tables, metrics and statistics.POC, specifications, development, packaging, deployment, internal documentation.Main achievements :- Cassandra wide rows / counters POCs- Spark / Spark SQL + Hive + Tableau / Spark streaming POCs- Redis / Aerospike POCs- MapReduce jobs (consolidations, metrics processing, exports, videos and users catalogs...)- google app engine modules refactoring (ETL, aggregations...)- ETL easy reprocessing- BigQuery automatic merge/aggregation per time period- website and player events pipelines (playback, ads, display, follow, repost...)- video streaming quality and advertising KPIs- Adobe SiteCatalyst import ETL (website metrics into BigQuery)- Datadog metrics export feature in hadoop and the ETL- Zendesk ETL (daily ticket metrics, tickets and users into BigQuery)- Datadog metrics ETL (daily network volumes into BigQuery)- video titles blacklisting analysis toolsTechnologies : scala, java, python, spark, hadoop, oozie, hbase, flume, nginx, cassandra, redis, google cloud platform (gcs, app engine, bigquery...), tableau

Worked on the dmcloud.net (now defunct) and dailymotion.com backends and various micro services (asynchronous tasks management, encoding, DRM, live...).POC, specifications, development, packaging, deployment, internal and public documentation.Main achievements :- search and related videos in dmcloud- dependency and recurrence in the asynchronous tasks management system- support for adaptive bitrate streaming- support for Microsoft PlayReady DRM to VOD and live streaming (from the media provisioning to the licence acquisition by the end user)- organizations (groups of users with centralized accounting/billing and rights management) in dmcloud- dmcloud streaming accounting refactor with parallelization- billing plans and payment (credit card with Stripe and check/transfer) in dmcloud- Mandrill integration for inbound/outbound email and Mailchimp for email lists in dmcloud- Zendesk integration in dmcloud- API ACL and validators refactor +automatic documentation in dmcloud- webhooks in dailymotion.comTechnologies : python, php, java, javascript, django, flask, mongodb, elasticsearch, gunicorn, nginx, unified streaming, ffmpeg, fabric