Mehdi Erraki

Designed and implemented real-time ingestion pipelines for robot telemetry data, including batch transformation and analytics layers on AWS. Designed a scalable data platform to support advanced analytics, business reporting, technical investigations, and API exposure. Developed a Python communication component for the robot’s embedded system, enabling cloud MQTT connectivity within a high-constraint environment following clean architecture principles. Led the data team and coordinated the company’s data governance strategy. Tech Stack: AWS IoT, Kinesis, Glue, Athena, PySpark, OpenSearch, Terraform, Gitlab CI

- [2022 - 2025] Lead Data engineer / Big data cloud architecture - Consultant @Aldebaran => Design and implementation of data engineering services ( AWS, Glue, Spark, Kinesis, Lambda, IOT, Athena, Opensearch, Security ) - [2021 - 2022] Senior Data Engineer Consultant @ Dalma Outstanding payment management service on top of AWS => AWS Glue, Lambda, API GW, Dynamo, PySpark, SAM - [2020 - 2021] Senior Data Engineer Consultant @ Bloom => design and implementation from scratch of a data engineering pipeline (Spark, AWS, Kafka, Delta) - [2018- 2020] Data Engineer Consultant @ IQVIA [healthcare] => architecture, design and development for a huge multi-country ETL migration into Cloudera Big data environment (Spark - Scala - Hadoop - Kudu - Impala)

Jellyfish (ex. Tradelab) is a multi-device programmatic buying platform that enables advertisers and agencies to connect with their best audiences through the use of large-scale data in real-time. In my team, we combine our technical big data expertise (kafka, spark, Google Cloud), and data science background to tackle a variety of missions : large scale data ingestion and processing, reporting, exploration, and data science. [BONUS] Efficient modelling for algorithmic bidding prospection.

End-of-studies internship within IT group department : - Design and POC of a data driven application for telco fraud detection based on CDR analysis (Hadoop Hortonworks, Hive, Pig, Spark) - Apache Spark : Advanced optimization and knowledge transfer (http://fr.slideshare.net/MehdiErraki/sparkintro1208) [BONUS] Large-scale unsupervised learning for Simbox fraud detection => 5 times more detected fraudsters

Participation in diffferent consulting and support missions for companies in their digital transformation.

Workshops animation, tools set up, communication and change management (Orange Sales Department).