#6: Scalable Low-Energy Molecular Conformer Generation with LoQI

Date and Time

• Date: Thursday, September 24, 2026

• Time: 9:00 AM – 10:00 AM EDT / 21:00 – 22:00 GMT+8 (20:00 – 21:00 GMT+7)

• Location: Online @ Zoom Webinar

Event Details

Join AI4Life, an independent open-science initiative and knowledge-sharing space for life science researchers, for a technical webinar exploring LoQI, a stereochemistry-aware generative model for scalable, low-energy molecular conformer generation recently accepted for publication in Nature Computational Science. The session examined how modern generative modeling overcomes the thermodynamic and stereochemical limitations of classical force fields, how LoQI samples low-energy conformational ensembles that approach quantum-mechanical accuracy, and how computational chemists can harness controllable parameters across discovery pipelines.

This webinar was hosted by Quynh Nguyen from the AI4Life Team. The format includes topic framing, a 20–25 minute technical keynote, an interactive live Q&A session, and a practical platform walkthrough. Across the expanding landscape of AI-driven drug discovery tools, the session will also showcase Vecura - an accessible, browser-based environment where researchers can run state-of-the-art open-source models like LoQI without coding experience or local GPU infrastructure, supported by an Academic Plan offering 6,000,000 free compute tokens for university researchers.

Speaker Information

Featured Speaker: Dr. Filipp Nikitin, Developer of LoQI

Filipp Nikitin is a developer of LoQI, a stereochemistry-aware generative model for scalable low-energy molecular conformer generation. He recently completed his PhD at Carnegie Mellon University, where his research focused on machine learning and generative modeling for molecular structure and chemistry. He is currently a Research Scientist at NVIDIA, working on AI methods for chemistry and molecular discovery. He is participating in this session in a personal capacity.

What You Will Learn

• The computational bottlenecks of conformational space sampling and thermodynamic accuracy in small-molecule modeling.

• How LoQI’s stereochemistry-aware generative architecture preserves chiral centers and geometric integrity.

• Benchmark evaluations showing how LoQI approaches quantum-mechanical (QM) fidelity and relative energy ranking at scale.

• Practical strategies and controllable parameter tuning for integrating LoQI conformers into downstream discovery workflows.

• A live demonstration of running LoQI on Vecura Workspace: browser-based, no-code execution with zero terminal setup, and how academic researchers can unlock 6,000,000 free compute tokens via the Academic Plan.

Register now to join the live session and bring your questions for the discussion!