Clément Tamines
Profile picture

Clément Tamines

Machine Learning Lead - PhD in Computer Science

Belgium

I’m a Machine Learning Lead at Proximus Ada, with a PhD in Computer Science, and practical expertise across Generative AI, LLM-based applications, and AI agents. I work at the intersection of strategy and execution, shaping technical direction while staying deeply hands-on in architecture and delivery.

My academic work at the University of Mons focused on strategy synthesis and verification in multi-agent games, including parity and generalized parity games, equilibria in non-zero-sum games, and related algorithmic questions, with publications and talks at international venues.

My path in industry has evolved from Data Scientist to ML Lead. Today, I lead and mentor machine learning engineers on cross-team initiatives, building shared AI foundations such as reusable components, tooling, and engineering guidelines that improve consistency, quality, and delivery speed. I also support data science teams directly with architecture and implementation for Generative AI, LLM applications, and agentic systems.

I’m driven by turning advanced AI into dependable production systems that create measurable impact and stand up to real-world complexity.

Interests

Equilibria in Algorithmic Game Theory Reactive System Synthesis Formal Verification Generative AI Agentic AI

Experience

Professional

Machine Learning Lead

Proximus Ada

Oct 2024 – Present
  • Built and managed a machine learning team, including strategy, hiring, and mentoring for cross-team initiatives.
  • Led a multi-month modernization of a shared repository of Generative AI assets, migrating it to GitHub, unifying code, libraries, and CI/CD, and supporting migration of 10+ use cases.
  • Evaluated AI agent frameworks, selected LangGraph, and defined reusable architecture patterns and tools to support adoption.
  • Implemented MLOps practices with MLflow, including tracing and evaluation workflows for LangChain and LangGraph applications.

ML Guild Lead

Proximus Ada

Feb 2024 – Sep 2024
  • Coordinated transversal machine learning initiatives with a part-time guild of data scientists.
  • Managed a shared Generative AI engineering repository, leading code reviews and feature development.
  • Organized weekly technical alignment sessions for 10+ data scientists to accelerate knowledge sharing and reuse.
  • Promoted common development practices and clearer technical communication across teams.

Generative AI Data Scientist

Proximus Ada

May 2023 – Sep 2024
  • Led and contributed to Retrieval-Augmented Generation (RAG) use cases, from technical analysis to production.
  • Introduced a shared Generative AI code repository, improving standardization and accelerating project delivery.
  • Developed deep expertise in LangChain and Azure to build custom LLM pipelines.
  • Worked closely with stakeholders to align scope, planning, and delivery expectations.

Data Scientist

Proximus Ada

Oct 2022 – May 2023
  • Worked on message classification pipelines using TF-IDF and transformer-based NLP techniques.
  • Compared clustering approaches on telecom signaling data for anomaly detection use cases.
  • Delivered stakeholder-facing analyses and dashboards to support interpretation and decision-making.

Academic

PhD Researcher

University of Mons

Oct 2018 – Sep 2022
  • Introduced and studied a theoretical framework for equilibria in multi-objective, multiplayer games.
  • Developed algorithms and tooling for verification and synthesis in games played on graphs.
  • Authored 5 peer-reviewed papers (3 conference, 2 journal).
  • Delivered 12 talks across conferences, workshops, seminars, and public events.

Teaching Assistant

University of Mons

Sep 2018 – Sep 2022
  • Teaching Assistant for graduate-level courses in algorithmics and bioinformatics.
  • Supervised projects and practical sessions, coaching students in implementation and problem-solving methods.

Education And Skills

Education

PhD in Computer Science

University of Mons

2022

University Certificate in Artificial Intelligence

University of Mons

2020

Master in Computer Science

University of Mons

2018

Bachelor in Computer Science

University of Mons

2016

Skills

Generative AI

  • RAG and LLM application design
  • LangChain, LangGraph
  • Azure OpenAI, Azure AI Search, ChromaDB

MLOps & Delivery

  • MLflow (tracing, evaluation, experiment workflows)
  • Airflow, Azure DevOps, Azure Pipelines
  • CI/CD practices for ML and LLM applications

Programming & Tooling

  • Python, Flask
  • Git, GitHub, uv
  • VS Code, PyCharm, JetBrains Suite, GitHub Copilot

Data, Visualization & Cloud

  • Matplotlib, Plotly, Seaborn, Streamlit, Dash
  • Azure (Compute Instances, Web Apps, Key Vault, Blob Storage)
  • Azure OpenAI, Azure AI Search

Leadership

  • Team coordination and mentoring
  • Strategic planning and vision setting
  • Interviewing and hiring

Communication & Collaboration

  • Stakeholder management and expectation setting
  • Public speaking, presentations, and technical communication
  • Knowledge sharing, meeting facilitation, and effective feedback

Publications

Journal Articles

The Reactive Synthesis Competition (SYNTCOMP): 2018-2021

STTT 2024

Swen Jacobs, Guillermo A. Pérez, Remco Abraham, Véronique Bruyère, Michaël Cadilhac, Maximilien Colange, Charly Delfosse, Tom van Dijk, Alexandre Duret-Lutz, Peter Faymonville, Bernd Finkbeiner, Ayrat Khalimov, Felix Klein, Michael Luttenberger, Klara J. Meyer, Thibaud Michaud, Adrien Pommellet, Florian Renkin, Philipp Schlehuber-Caissier, Mouhammad Sakr, Salomon Sickert, Gaëtan Staquet, Clément Tamines, Leander Tentrup, Adam Walker

Stackelberg-Pareto Synthesis

ACM TOCL 2024

Véronique Bruyère, Baptiste Fievet, Jean-François Raskin, Clément Tamines

Conference Papers

Pareto-Rational Verification

CONCUR 2022

Véronique Bruyère, Jean-François Raskin, Clément Tamines

Stackelberg-Pareto Synthesis

CONCUR 2021

Véronique Bruyère, Jean-François Raskin, Clément Tamines

Partial Solvers for Generalized Parity Games

RP 2019

Véronique Bruyère, Guillermo A. Pérez, Jean-François Raskin, Clément Tamines

Preprints

The Reactive Synthesis Competition (SYNTCOMP): 2018-2021

CoRR

Swen Jacobs, Guillermo A. Pérez, Remco Abraham, Véronique Bruyère, Michaël Cadilhac, Maximilien Colange, Charly Delfosse, Tom van Dijk, Alexandre Duret-Lutz, Peter Faymonville, Bernd Finkbeiner, Ayrat Khalimov, Felix Klein, Michael Luttenberger, Klara J. Meyer, Thibaud Michaud, Adrien Pommellet, Florian Renkin, Philipp Schlehuber-Caissier, Mouhammad Sakr, Salomon Sickert, Gaëtan Staquet, Clément Tamines, Leander Tentrup, Adam Walker

Pareto-Rational Verification

CoRR

Véronique Bruyère, Jean-François Raskin, Clément Tamines

Stackelberg-Pareto Synthesis (Full Version)

CoRR

Véronique Bruyère, Jean-François Raskin, Clément Tamines

Partial Solvers for Generalized Parity Games (Full Version)

CoRR

Véronique Bruyère, Guillermo A. Pérez, Jean-François Raskin, Clément Tamines

Talks

2022

On Pareto-Optimality for Verification and Synthesis in Games Played on Graphs (in French)

public PhD defense

Mons, Belgium

Pareto-Rational Verification

CONCUR 2022

Warsaw, Poland (online)

Pareto-Rational Verification

Highlights 2022

Paris, France

2021

Stackelberg-Pareto Synthesis

Journées annuelles du GT Vérification du GDR IM

Gif-sur-Yvette, France

Stackelberg-Pareto Synthesis

Highlights 2021

Aachen, Germany (online)

2020

Formal Verification Using Game Theory (in French)

Séminaire Jeunes

Mons, Belgium

2019

Partial Solvers for Generalized Parity Games

Highlights’19

Warsaw, Poland

Partial Solvers for Generalized Parity Games

RP’19

Brussels, Belgium

Partial Solvers for Generalized Parity Games

MoRe’19 (LICS’19 workshop)

Vancouver, Canada

Your Turn to Play! (Popular Science talk on Game Theory)

Math & Science Days’19

Mons, Belgium

Projects

Clément Tamines preview

SPORE (Symbolic Partial sOlvers for REalizability)

A prototype symbolic implementation of partial solvers for (generalized) parity games applied to LTL realizability.

Clément Tamines preview

Forest fire detection using CNN

Using convolutional neural networks (CNN) to detect the presence or the start of a forest fire in an image. This model could be applied to detect a fire or a start of a fire from (aerial) surveillance footage of a forest.