Anssi Mäkiniemi
Head Of Research Data Services
London, England, Finland
Summary
🤩
Rockstar🎓
Top SchoolAnssi Mäkiniemi is a data and digital health leader with over a decade of experience shaping R&D data strategy and operational delivery at Orion Pharma, now heading Research Data Services in London. He combines hands-on engineering instincts—evidenced by substantial open-source contributions to reinforcement learning projects like stable-baselines3 and ViZDoom—with executive responsibilities for budgeting, resourcing, partnerships and regulatory-compliant processes. His background spans GxP validation, medical device software releases and enabling scalable data ecosystems that accelerate scientific discovery. Comfortable at the intersection of Data Science, Clinical Operations and product teams, he translates research-grade ML work into compliant, production-ready solutions. Notably, his GitHub work includes audio integration for RL environments and core algorithm refactors, reflecting a practical focus on performance and reproducibility. He holds an MSc in Chemistry from the University of Turku and brings a rare blend of lab-based scientific experience and applied machine learning engineering.
10 years of coding experience
9 years of employment as a software developer
MSc, Chemistry, MSc, Chemistry at Turun yliopisto - University of Turku
Finnish, English, Swedish
Github Skills (30)
algorithm10
minecraft10
algorithms10
pytorch10
ppp10
python10
testing10
windows10
gymnasium10
machine-learning10
data-structure10
reinforcement-learning10
deep-learning10
tensorflow10
data-structures10
Programming languages (9)
Github contributions (5)
openai/Video-Pre-Training
Jun 2022 - Aug 2022
Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 5 reviews, 16 commits in 2 months
Contributions summary:Anssi contributed to an AI project focused on video pre-training. They added usage instructions, fixed a typo, and integrated initial Inverse Dynamics Model (IDM) code. Furthermore, the user developed code to compare the predicted actions from the IDM with true data and included example code for behavioral cloning.
actvideospretrainingvideovpt
minerllabs/minerl
Jun 2020 - Jan 2023
MineRL Competition for Sample Efficient Reinforcement Learning - Python Package
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:3 releases, 47 reviews, 102 commits in 2 years 7 months
Contributions summary:Anssi primarily focused on addressing platform-specific issues and implementing new features within the MineRL environment. This included debugging and fixing issues on Windows, such as build problems and multiple instance conflicts. They also added new functionalities related to chat actions and Minecraft commands, expanding the interactive capabilities of the environment. Furthermore, the user contributed to documentation and addressed minor typos to enhance usability.
pythonreinforcement-learning
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.