Aschwin F. Schilperoort

dsm-firmenich

Senior Scientist - Data Science • Feb, 2025 — Present

• Further improved Bayesian optimization framework by integrating multi-fidelity modeling (botorch, gpytorch) to learn patterns across fermentation scales. Applied to 4 product development projects, proposing optimal blends for yogurt and cheese properties. One design unlocked a major U.S. dairy producer, generating €3M annual revenue.
• Started a new data science team for a specific business unit, hiring 2 data scientists, defining the operating model, and shaping the data science strategy for all yogurt and cheese R&D projects.
• Currently researching genomics-informed Bayesian optimization, leveraging microbial genomic data to guide strain selection and improve starter culture performance through more efficient optimization.

dsm-firmenich

Scientist - Data Science • Oct, 2023 — Feb, 2025

• Improved Bayesian optimization framework by moving to botorch, enabling multi-objective and batch optimization for culture optimization problems. Delivered optimized yogurt and cheese starter cultures, including a zero-sugar yogurt and a high-texture, low post-acidification variant.
• Setup and implemented a Laboratory Information Management System (Sample Manager) for dairy fermentation R&D, adopted by 5 technicians to capture data FAIRly. Connected experiment data into a simple analysis pipeline with Python, SQL, and PowerBI, reducing manual reporting and enabling cross-project insights.
• Analytics and DOE support for commercial product development that contributed to Delvo One winning “Digital Transformation Innovation of the Decade” at Gulfood Manufacturing.

dsm-firmenich

Senior Associate Scientist - Data Science • Feb, 2022 — Oct, 2023

• Lead data scientist/engineer for the Co-Creation Platform supporting global dairy product development.
• Applied Bayesian optimization with GPy to microbial strain combinations, enabling a 10-day shelf life increase in yogurt. Novel designs resulted in the Delvo One culture range, unlocking a major Latin American dairy client (~€1M annual revenue).
• Designed and deployed ML pipelines on AWS (Lambda, Batch, S3, RDS) with SharePoint API integration, automating recipe design workflows for yogurt and cheese cultures.

dsm-firmenich

Associate Scientist - Data Science • Nov, 2020 — Feb, 2022

• Automated data collection from lab robots (Python + PowerShell → PostgreSQL + FastAPI), eliminating manual file handling.
• Built CNN in PyTorch to quantify fungi growth in microtiter plates; automated via Airflow, visualized in Spotfire for scientists.
• Developed full-stack Flask app for strain engineering (frontend, backend, CI/CD), enabling genotype-phenotype tracking for future R&D scale-up.
• Established FAIR data storage in CKAN for baking & alt-protein R&D; integrated with Spotfire dashboards, reducing Excel dependence.

Xccelerated (part of Xebia)

Data Scientist • Oct, 2019 — Oct, 2020

Followed an intensive bootcamp on delivering production grade data science solutions with pandas, scikit-learn, and Git and weekly trainings, applied advanced methods (Bayesian models, PySpark, Docker, CI/CD, Neural Networks). Worked four days at client, dsm-firmenich.

• Automated weekly NMR data workflows with Apache Airflow and PostgreSQL. Eliminated manual Excel email exchanges and saved hours of technician time for ~6 active lab users.
• Built FAIR data storage for proteomics workflows in CKAN and PostgreSQL, improving findability and access for ~10 scientists and reducing time wasted searching for correct datasets and cleaning excel files.

Unilever (Datlinq Consultant)

Data Scientist • Feb, 2018 — Sep, 2019

• Built sales performance dashboards in Tableau and Power BI for the Out of Home sales teams of Lipton, Ola and Ben & Jerry’s, giving managers a clear view of results across regions.
• Used SQL and Python (pandas, scikit learn) on Point of Sale data to create regression models for estimating revenue potential and classification models to predict product conversion.
• Combined revenue and conversion models to calculate ROI per sales visit and designed optimized sales routes, helping the sales team focus on the most valuable locations.

Landscape - Increasing data impact

Accountmanager • Apr, 2017 — Jan, 2018

Landscape is a data science company that was still in its early stage. My role was to bring in new clients and look after existing ones. I set up a clear sales pipeline and sales procedures that are still used today. I was responsible for all new business during my time there and brought in the company’s first €50,000 in revenue through cold acquisition.

SQPeople

Salesmanager • Mar, 2013 — May, 2017

Part-time position while studying. Responsible for the sales efficiency of a sales establishment of approximately 40 people. Training sales agents, via self-made workshops and training on location. Working logistics and assisting the general manager.

Albert Heijn

Stock clerk • Jul, 2008 — Aug, 2012

Part-time position while studying. Responsible for stock management, inventory control, and customer service.

Leiden University

Bachelor of Science, Mathematics • 2009 — 2016

The Mathematical institute covers a broad range of topics, with a focus on Algebra, Geometry and Number Theory; Analysis and Dynamical Systems; Probability Theory; and Statistics. My courses were mostly focussed around Decision Theory and Statistics.

• Thesis with the title 'Polya Tree Priors', 2015, math.leidenuniv.nl

Technical University Delft

Propedeutic Diploma, Civil Engineering • 2008 — 2009

Civil engineering is concerned with the development, design, construction, production and management of the physical infrastructure required to safeguard the safety, health, business activity and sustainability of our society.

Alfrink College

VWO Diploma • 2002 — 2008

scikit-ferm

Founder & Primary Developer • 2025 — Present

Scikit-Ferm is an open-source Python library for fermentation modeling and optimization, built on top of scikit-learn and other scientific libraries.

Improvlib

Co-Founder & Organizer • 2024 — Present

Improvlib is a community-driven library for improvisational theater resources, including games, exercises, and performance tips.

Game of Amazons

Contributor • 2023 — Present

Game of Amazons is a strategy board game that involves tactical movement and area control, with a focus on player interaction and dynamic gameplay.

• Chess
• Tennis
• Music (guitar)
• Improvisational theater