SCIENTIST

BeerseIPharmaceutical Sector I Manufacturing

Job description

For one of our pharmaceutical clients regio Kempen we are currently looking for a SCIENTIST

We are looking for a person that can handle convergence analytics and architecture optimization for: 

  • Gated message passing neural networks (mechanistic interpretation of chemical reactions). The current implementation, like ELECTRO, are too slow and need to be changed to ensure they utilize the full power and distribution of a DGX2 GPU architecture as well as preprocessing costly compound annotations. Those concepts should be consolidated across PyTorch deep learning architectures to ensure a leveraging of implementing caching solutions.
  • Enablement of Large-scale monte-carlo-tree search (MCTS) by identifying options to store chemical building blocks and chemical reactions and consume them efficiently for MCTS optimization flows.
  • Latent space integration for deriving a goal-orientation via conditional generative deep learning architectures. This enables the exploitation of existing data assets. Retrosynthesis integration via MCTS by enabling business and R&D relevant execution policies to ensure business relevance. Reinforcement-learning for forward synthesis to aim at exploring novel spaces, which might not be reachable by a synthesis-driven sampling alone, e.g. automated chemistry. Thus, the goal is to maximize the goal-orientation and focus on the chemical synthesis, reactions, and building blocks enabling this space.
  • Compound-level integration in a way to create smooth latent spaces enabling a goal-directed sampling for compound optimizations and suggestions. The very same concepts should then be deployed on a chemical reaction level to ensure chemical reactants can lead to the products supporting such a goal orientation. 

In addition, you also need to be able to apply the above in projects that utilize all of the above, e.g. projects to find novel inhibitors for the inflammasome.

Profile
  • You have a Master degree in a scientific field
  • You possess 4 to 6 years' working experience
  • You have experience with Deployment of Kubernetes/docker deployments as single deployments or flows e.g. Argo (basis of kubeflow)
  • You have knowledge of Data and model persistency via DVC, and Model serving as REST or GraphQL
  • You have experience with Jupyter Enterprise Gateway and conda and dockers, e.g. based on Nvidia base images
Organisation and department description

Select Projects is part of the Select group, a collection of HR companies that put people first because Quality is Performance by Genuine People. Select Projects provides technical support to clients on a project basis. Our 220 employees in Belgium and the Netherlands support companies with their projects in various domains: Life Sciences, (petro) chemical, Construction and Rail. From our offices in Antwerp and Rotterdam we optimally support our experienced specialists so that they can deliver high quality to our customers at all times.

Offer You get an attractive salary package with interesting fringe benefits. A company car, meal vouchers, group insurance and mobile phone subscription are possible. We ensure that you end up in a pleasant working environment with nice colleagues and opportunities to develop and further develop. Together with you, we map out a training course tailored to your future plans and needs.
Any questions?
Marlies Collart Consultant Life Sciences Select Projects +32 3 203 50 80
SCIENTIST Vacancy information
77526
Life Science
Project Sourcing
What is my travel time?

Applying at Select Projects

Applying for a job on our website?

  • You apply for a job at Select Projects
  •  
  • You will usually receive a message within one workday
  •  
  • We get acquainted with you in a (phone) discussion
  •  
  • If it is a good fit all around then we will make a proposal
  •  
  • We're going to work. together. Welcome to Select Projects
  •  
  • Your coach provides personal career guidance
  •  
  • New projects and experiences keep you on your toes
More information on our application procedure