Webinar → Accelerating the Lab-in-the-Loop

​AI can generate thousands of candidate molecules in hours. The wet lab still takes 4–6 weeks per round, and that limits how many turns a discovery programme gets. Most teams running lab-in-the-loop use AI to improve their odds on each round. Few have managed to also increase the number of rounds they can run. Getting both, better molecules from every loop and faster loops, is exactly what this webinar is about.

​Moritz Freidank (AI Engineering Director at Visium) will walk us through each step of the loop, from generative design and property prediction to active-learning selection and wet-lab validation, and share what has worked in real programmes. He'll cover how generative models, developability scoring and ranking work together to balance exploiting your best candidates with exploring to improve the models. He'll also show how pre-trained models get results from only tens to hundreds of labelled data points per programme, and why a model trained on one programme often doesn't carry over to the next. You'll see a real case study: a VHH programme that went from 3% to 92% specific binders, and from µM to nM binding strength, in two lab rounds. Finally, he'll look at how to run more rounds, with automated lab-data integration and assay-image review, plus where closed-loop, agent-to-instrument setups are heading.

​Expect a practical session for R&D leaders, computational and medicinal chemists, protein engineers, and data science teams in biopharma. Bring your questions, since there will be time for live Q&A with Moritz. We'll meet online on Zoom.

​About Visium
Visium is a leading Swiss AI & Data company that builds the AI systems trusted to run the core workflows of Life Sciences & Specialty Chemicals, so the world's most vital sciences move faster.

​Follow us
→ Website
→ Newsletter
→ LinkedIn
→ Youtube