Cas Wognum

Cas Wognum

  • Montréal, Canada
  • August 01, 1997

One man, many hats!

In my work, I strive to collaborate with an inspiring group of people on projects that I am proud of. Within a team where people treat each other as equals, I want to be challenged and use all of my capacities to have a positive impact on society. In pursuit of these values, I have worn different hats. Whereas film-making was my first passion, I’ve more recently become fascinated by the mystery that the human mind proves to be and by the possibilities that the field of artificial intelligence promises.

Contact

RECENT Experience

Valence Labs (Recursion)

Associate SCIENTIST

2023 - now

Following the acquisition of Valence Discovery by Recursion, Valence Labs continues as a semi-autonomous R&D engine.

Valence Discovery

ML SCIENTIST

2021 - 2023

After a succesful graduate internship, I was hired to bring the latest advances in machine learning for chemistry to the Valence platform.

Orbisk

ML ENGINEER

2019 - 2021

Not only does Cas add a lot of value with his expertise on deep learning and computer vision, but also his professional and structural attitude makes him a pleasure to work with. His attention to detail, ability to work autonomously and when necessary act as a team player is a big relief in the sometimes hectic environment of a startup. Being there from the start, Cas has been an integral part in developing our team and team spirit to where it is today.

Bart van Arnhem, co-founder and CTO at Orbisk

EDUCATION

Utrecht University

Artificial
intelligence

MSc. / 2019 - 2022 / Cum Laude

Learnt a multi-disciplinary and history view of Artificial Intelligence, while remaining up-to-date with the latest developments in the field.

Utrecht University

Game and media technology

MSc. / 2019 - 2022 / Cum Laude

Learnt to build and simulate virtual worlds. Through extensive optimization, generated photo-realistic images of these worlds in real-time. 

Eindhoven University of Technology

Computer Science
and engineering

BSc. / 2016 - 2019 / Cum Laude

Gained a deep understanding of how computers operate and learned to effectively design and collaborate on large-scale software projects.

Science

Real-World Molecular Out-Of-Distribution: Specification and Investigation.
Tossou, P., Wognum, C., Craig, M., Mary, H., & Noutahi, E. (2023).
paper | code | blog | video

MoDTI: Modular Framework For Evaluating Inductive Biases in DTI Modeling
Eyono, R. H., Tossou, P., Wognum, C., & Noutahi, E. (2023).
paper
Challenging the i.i.d. assumption in molecular modeling (MSc. thesis, 2022)
My thesis on Molecular Out-of-Distribution Generalization. 
 thesis | defence

→ See more on Google Scholar

Code
Polaris (2024, open-source)
Community-run, industry-backed benchmarking platform for drug discovery.
code
Graphium (2023, open-source)
Advanced library to more easily scale molecular Graph Neural Networks.
code | blogpost
Molfeat (2023, open-source)
Unifies several molecular representations behind a user-friendly interface.
code | blogpost | website

→ See more on Github

Film
Molecular Out-of-Distribution (2023, English, animation)
Scientific animation to convey the high-level ideas of a recent article.
video
Geen Bereik (2015, Dutch, very short film)
After “Sukkel voor de liefde”, this was an intentionally minimalistic project to experiment with what we had learnt.
video
Sukkel voor de Liefde (2014, Dutch, short film)
Made in just 30 days as part of a national sex education campaign.
video | application

→ See more on Youtube

Miscellaneous
M2D2 community (founding member)
The M2D2 community hosts various initiatives to bring together scientists in the field of AI-enabled drug discovery.
website

November 22, 2024