Work experience

Ten years of data science across consultancy, universities, research institutes and international organisations, in New Zealand, France and Switzerland. Most recent first.

Download my CV See the work

Founder and Data Consultant

DataSphere Analytics · Montpellier, remote

2024 - present
  • Design, build and ship interactive Shiny applications end to end for research institutions, NGOs and government clients, from scoping to deployment and handover: FINS Explorer and SAFE for the University of Zurich, DateBack for IRD and CNRS, and a survey dashboard with continuous integration for Citizens for Europe.
  • Built, with Yan Holtz, an automated Quarto pipeline for the Government of Vanuatu that turns a cyclone scenario into a costed damage and response assessment (public demo).
  • Led a working group's data pipeline at the Global Fishing Watch Mediterranean Data Lab (Rome, October 2026): one GitHub repository, a live Quarto report and slides that update themselves.
  • Wrote data visualisation tutorials for the R Graph Gallery (chord diagram, Sankey with highlight, refugee flows) and publication-ready network figures for the University of Chile.
  • Published a reusable, branded Quarto report template and a series of open analyses of satellite and vessel-tracking data, listed on the portfolio.
Logos of DataSphere Analytics clients and partners
RShinyQuartoPythonGit and GitHubClient delivery

Data Scientist, Postdoctoral Researcher

IRD, French National Research Institute for Sustainable Development, BRIDGES programme · Montpellier

Jan - Jul 2026
  • Built an LLM-assisted extraction pipeline in Python with the Anthropic Claude API, pulling 27 structured fields from a corpus of more than 3,000 documents.
  • Validated it at scale against blind manual coding: inter-rater agreement (Gwet's AC1), cluster bootstrap confidence intervals and McNemar's test.
  • Delivered the screening and analysis system as a documented public GitHub repository, structured for reuse by the client team.
PythonLLMsAgreement statisticsReproducibility

Data Scientist, Postdoctoral Researcher

University of Zurich, Pimiento Research Group · remote from Montpellier

2025
  • Led a global conservation prioritisation of sharks and rays with CAPTAIN, a reinforcement learning algorithm in Python, optimising across about 1,000 species and three conservation criteria. Available as a preprint.
  • Trained a random forest on paired AIS and Sentinel-1 radar data to estimate fishing effort where vessels stop broadcasting (repo).
  • Ran a gap analysis of the global marine protected area network against a spatially explicit null model, with quantile and beta regression.
  • Built the FINS and SAFE Shiny applications for the group's open datasets (FINS paper).
Machine learningReinforcement learningSatellite dataR and Python

Postdoctoral Researcher and Scientific Coordinator

IUCN SSC Shark Specialist Group, Important Shark and Ray Areas · remote

2023 - 2024
  • Coordinated more than 60 scientists across international institutions to design, write, review and publish the regional compendia for the Mediterranean and Black Seas and the Western Indian Ocean, to deadline.
  • Co-organised and delivered two expert workshops with more than 250 participants in total, in Thessaloniki and Durban.
  • Led an analysis of protection shortfalls in the Central and South American Pacific with text mining, mixed effects models and spatial overlaps, published in Marine Policy.
CoordinationWorkshops and trainingStakeholdersSpatial analysis

PhD in Macroecology (statistical modelling)

MARBEC, University of Montpellier, degree from Université Claude Bernard Lyon 1

2020 - 2023
  • Compiled, cleaned and modelled a 30-year database spanning 70 sites and more than 100 variables (biodiversity, climate, land use and habitat), building the full workflow in R.
  • Assembled and managed a team of 15 scientists across three continents.
  • Published three first-author papers, including Global Change Biology and Ecography (thesis).
Statistical modellingLarge datasetsRTeam leadership

Research Scientist

CESAB, French Foundation for Research on Biodiversity · Montpellier

2021 - 2022
  • Built end-to-end R pipelines to format, clean and analyse a database of more than 9 million rows linking human population growth to protected area establishment across the United States over two centuries.
  • Contributed to the PARSEC project on reproducible, collaborative data science.
Data pipelinesRCensus and land-use data

Research Scientist

Ifremer, French Research Institute for Exploitation of the Sea · Sète

2019 - 2020
  • Supported a coastal ecological monitoring network with data cleaning, analysis and reporting, and ran its fieldwork and laboratory campaigns.
  • Led a mixed-method project: a trait database, multivariate statistics and generalised models, with weekly reporting to an expert panel (paper).
  • Hired and supervised an MSc student.
Monitoring dataMultivariate statisticsSupervision

Research Assistant

NIWA, National Institute of Water and Atmospheric Research · Hamilton, New Zealand

2017 - 2019
  • Statistical programming, data processing and multivariate analyses for three research groups in marine and freshwater science.
  • Species distribution modelling of New Zealand cetaceans and marine spatial classification with machine learning, for Fisheries New Zealand and the Department of Conservation (report, paper).
  • Led a first-author paper in Science of the Total Environment and co-supervised an MSc student.
Machine learningSpecies distribution modelsData visualisation

Education

PhD in MacroecologyUniversité Claude Bernard Lyon 1, research at the University of Montpellier
2020 - 2023
MSc Dynamics of Aquatic EcosystemsUniversity of Pau and Pays de l'Adour, thesis at Massey University, New Zealand
2014 - 2016
BSc Ecology and Organismal BiologyUniversity of Montpellier
2012 - 2014