NVIDIA-Accelerated AI and Autonomous Labs for Therapy Design

An overview of how Lila Sciences connects AI to an autonomous laboratory accelerated by NVIDIA, using a therapy design goal for B-cell lymphoma as the example. The described workflow starts with scientific reasoning, moves through computational virtual experiments, and continues into robotic testing with results fed back into the next cycle.

■ Workflow and tools
- Lila Iris builds an experimental plan from the scientist’s goal using NVIDIA Nemotron and NVIDIA Agent Toolkit
- A computational dry lab runs virtual experiments with domain-specific models built using NVIDIA BioNeMo
- AI Science Factory robots make and test candidates, with NVIDIA Isaac and Omniverse for robotic development and simulation and Jetson and Thor for compute
- Experimental results return to Iris to inform subsequent designs and experiments

■ Scope
- B-cell lymphoma therapy design illustrates a closed-loop AI science workflow
- Covers the conceptual workflow and named NVIDIA technologies; does not present detailed protocols, performance metrics, pricing, or timelines

Intended for viewers interested in AI for life sciences, laboratory automation, and NVIDIA-enabled scientific workflows. It provides a high-level view of how AI planning, simulation, and robotics can be connected in an iterative research process, helping viewers understand the roles of the named tools. Viewers can use the overview to identify which components warrant deeper technical review.