Data Engineer, ML Platform

Hybrid (4 days/week) · Bordeaux (France) · Mid · Full-time

Want offers like this one?

Create your free PyHires profile: you will see how well you match each offer and get a weekly email with the ones that fit you.

Create my profile

Overview

Work mode
Hybrid (4 days/week)
Location
Bordeaux (France)
Engagement
Full-time
Seniority
Mid
Languages
French, English
Open to candidates from
EU countries only
Company
Mirakl
Posted
Nov 19, 2026

About the role

Mirakl builds the operating system for marketplace commerce, used by 450+ marketplaces. In the Data Experience squad of Mirakl Labs, part of a data team of 70 specialists, you will build and scale the ML infrastructure behind Mirakl's AI products, from real-time recommendation inference to LLM fine-tuning pipelines. This is a permanent position (CDI) in Bordeaux, with 4 days on site per week. What you'll do - Design a performant, robust and scalable ML platform for training, evaluation and inference, while keeping costs under control - Maintain and improve ML production pipelines and real-time inference at scale (recommendations, 1M+ LLM calls a week, async content moderation) - Partner with Data Scientists from kickoff to production, setting best practices, guardrails and CI/CD improvements with the SRE team - Manage technical debt, systematise version upgrades and upskill the team on MLOps and data engineering - Contribute to AI projects such as LoRA adapters for open-source LLM fine-tuning, in-house computer vision models and large-scale product data transformation What you bring - 2+ years in a machine learning and/or data engineering environment - A track record of shipping ML applications at scale (NLP, image processing or classical ML) - A software development background with a strong appetite for data science and ML - Proficiency in Python and the Kubernetes/Docker ecosystem - Solid experience with cloud environments (mainly AWS) and Apache frameworks - Fluent French (C2) and good communication and mentoring skills - Stack: Python, Databricks, AWS, Kafka, Airflow, Prefect, MLflow, TensorFlow, Triton, PyTorch, vLLM, KServe, Delta Lake, Kubernetes, SQL, Terraform and Spark What they offer - A permanent contract (CDI) in Bordeaux - Hybrid work: 4 days on site, 1 remote day per week - Cutting-edge AI projects in a member of the French Tech Next40

Requirements

Based on:Python

Must-have

  • AWS
  • Apache Airflow
  • Apache Kafka
  • Apache Spark
  • Docker
  • KubernetesEssential
  • Prefect
  • PyTorch
  • SQL
  • Terraform

About Mirakl

Mirakl, founded in 2012, builds an operating system for marketplace commerce: an enterprise marketplace and dropship platform, AI-powered multichannel selling, retail media and agentic commerce infrastructure (Mirakl Nexus). With dual headquarters in Boston and Paris, it serves 450+ marketplaces and more than 100,000 sellers for brands such as Macy's, Decathlon, Carrefour, Asos and Airbus Helicopters, and has 750+ employees in 9 offices.

Stack

  • Python
  • AWS
  • Apache Airflow
  • Apache Kafka
  • Apache Spark
  • Docker
  • Kubernetes
  • Prefect
  • PyTorch
  • SQL
  • Terraform

Company