Roland Macheboeuf

Data, machine learning and software craftsmanship — from prototype to production.

ML and data engineer with 4+ years shipping to production. From a marine insurance SaaS — taken from zero to profitability — to data products for large enterprises (TotalEnergies, IPSEN), I turn business needs into reliable, maintainable systems that people actually adopt.

Portrait of Roland Macheboeuf

Expertise

What I work on

Before writing a single line of code, I seek to understand the actual need: I ask a lot of questions, scope the problem, then propose a solution that fits it — rather than chasing the latest trend.

Data engineering & Data Mesh architectures

  • Data platforms in production on Databricks and cloud (AWS, Azure)
  • Data Mesh, Spark/DBT/Airflow pipelines, governance and FinOps
  • Shared frameworks (inner-sourcing) to unify practices and governance across domains

Data science & AI

  • Computer vision, deep learning, risk models on unstructured data
  • Solid mathematical foundations (Centrale Lille + Master's degree in mathematics)
  • Continuous learning: reading papers, evaluating recent models (generative AI, LLMs)

Machine learning in production & MLOps

  • From prototype to a deployed, maintained model
  • Testing, monitoring, CI/CD, cloud and edge deployment
  • Models designed to withstand production constraints and last

Software craftsmanship & technical leadership

  • Domain-Driven Design and hexagonal architecture: business logic isolated from infrastructure
  • Robust, maintainable code that mirrors the business domain
  • TDD, code reviews and reusable templates that upskill teams

Career

From first models to production

  1. OCTO / Accenture · Data consultant

    2024–2026

    Data Mesh architectures in production at large enterprises (TotalEnergies, IPSEN, Streem Group).

  2. Trapil · Computer Vision ML Engineer

    2023–2024

    Defect detection on ultrasonic images of hydrocarbon pipelines, deployed at the edge.

  3. Meetrisk · ML Engineer & Full Stack

    2022–2023

    Marine risk rating SaaS platform, taken from zero to profitability.

  4. Amadeus · ML Researcher

    2018

    Deep learning (LSTM, CNN) for caption generation and image ranking, in a B2B context.

  5. View the full career path

Articles

Recent articles

View all articles
July 2026Data mesh

Data mesh in motion: 6 scenarios to understand the roles and the tooling.

Who is responsible for what, with which tooling blocks, and how data flows from creation to retirement: six practical scenario guides to make the data mesh concrete — no vendor pitch.

2025Article

How can Domain-Driven Design and Hexagonal Architecture improve Data Product development in practice?

How DDD and hexagonal architecture can provide concrete structure for data product development — article on the OCTO blog, with a demo repository.