BY
7 min read
About the episode
This is an unedited live walkthrough recorded with the computational biology team at a Series B biotech (name withheld at their request). Starting from a set of target binding properties, they ran 500 candidate screens, reviewed the top 20 with explainability reports, and shortlisted 4 for lab validation - in a single afternoon session.
We kept the whole thing in. Including the one moment where Marcus had to explain why candidate #7 ranked higher than the team expected.
Chapters
0:00 | Introducing the research brief (target properties, constraints) |
3:45 | Uploading parameters and running the initial screen |
9:20 | Reading the ranked candidate list with confidence scores |
16:10 | Drilling into the explainability report for top 3 candidates |
24:30 | The team pushes back on candidate #7 - live discussion |
31:00 | Final shortlist of 4 selected for lab validation |
35:15 | What happens next - handing off to the lab queue |
