Main
Willian Vieira PhD
Applied ML scientist with a PhD in quantitative ecology. I build statistical models, production-minded ML pipelines, and reproducible systems that turn noisy historical data into validated algorithms and decisions people can trust.
Data Science & Engineer Experience
Data Analyst & developer
Habitat, Montreal, Canada
N/A
2025 - 2024
(1 yr 9 m)
Built production-minded R/Python analytical pipelines that turned messy real-world project data into reusable, validated model inputs and decision-ready outputs | Prototyped and iterated data products with historical/project data, using validation checks and stakeholder feedback to improve reliability and usefulness | Designed metadata-driven ETL and provenance workflows so datasets, definitions, analyses, and handoffs remained traceable, reproducible, and inspectable | Led the transition to a Unix-based production environment using Docker, CI/CD, automated testing, documentation, and collaborative development practices | Designed cloud/Azure-oriented data infrastructure that made heterogeneous analytical datasets easier to retrieve, validate, and reuse
PhD Research
Integrative Ecology Lab, Sherbrooke, Canada
N/A
2024 - 2017
Developed Bayesian hierarchical models, simulation workflows, and historical-data evaluation routines to forecast forest dynamics from noisy, biased, incomplete real-world data, with explicit uncertainty propagation and model comparison | Ran thousands of simulations on HPC infrastructure () to test algorithmic assumptions, compare model behavior, and scale analyses beyond local compute | Built reproducible pipelines for continental-scale geospatial and environmental data, harmonizing climate, satellite, and field-inventory datasets into modeling-ready inputs | Developed open-source software libraries (, ) for demographic modelling with version control, automated testing, documentation, and reusable workflows | Published a technical methods book documenting the full computational pipeline, along with one peer-reviewed publication and two preprints (, )
Biostatistician
Environment and Climate Change Canada - Quebec, Canada
N/A
2022 - 2020
(part-time)
Developed a cost-aware probability sampling protocol to improve the spatial representativeness of boreal bird surveys in Quebec | Led R&D on spatial bias-correction methods for large-scale ecological monitoring data, designing an approach later adopted by other provinces | Engineered a fully automated and reproducible data pipeline with version-controlled workflows, automated documentation, and stakeholder-ready analytical outputs
Education
PhD, Ecology
Université de Sherbrooke - Sherbrooke, Canada
N/A
2024 - 2017
How climate, competition, and forest management shape the limits of tree species distributions: from individuals to metapopulations
Masters 2, Agroecology and Resource Management
Bordeaux Sciences Agro, Bordeaux, France
N/A
2016 - 2015
Modelling the dispersion of weed species in agricultural landscapes