Research

Using Agents to Parameterize Molecular Dynamics Biomolecules

From Bayesian Optimization to LLMs, exploring the potential of AI agents for parameterizing molecular dynamics biomolecules.

Evaluation Driven Agentic Design: An agentic materials research use case

An evaluations-first approach to building reliable agentic systems for scientific discovery using Google's Agent Development Kit.

森 forest, a knowledge layer for your Slack workspace

A self-hosted Slack bot that passively captures files and links shared in your workspace, uses an LLM to organize them into a virtual folder tree, and lets you browse or search them with a simple mention command.

Using Graph Convolutional Networks to generate synthetic data in Unity

Leveraging Graph Convolutional Networks to automatically generate synthetic 3D data in Unity that closely matches a desired target dataset.

From Blender to Transformers: Fine-Tuning DETR on Synthetic Data

A walkthrough of generating synthetic object detection data in Blender with sdg-engine and fine-tuning a DETR model, with full training curves and qualitative results.