A modular pipeline for Agentic Neural Architecture Search combining LLMs and NAS.

Open-source deep neural network library to accelerate inference on AMD Zen processors.

A unified sim-and-real benchmark for comprehensive evaluation of generalist robot manipulation policies.

A model for verifiable TLA+ specification synthesis via preference-optimized LoRA.

Latent Generative Temporal Augmentation model for time series augmentation.

A scalable framework to balance training measures for discretized physical simulations.

An agent-first full-stack realtime framework.

A topology-preserving representation method based on high-capacity Self-Organizing Maps.

A browser video editor specifically designed to be controlled by AI agents with zero dependencies.

A competitive programming benchmark for evaluating the coding capabilities of LLMs.

Foundation model for end-to-end time series classification via in-context learning.

A synthetic AI-lab codebase designed for testing AI control and distributed attacks.

A plug-and-play adapter for aligning robot proprioception with vision for generalist manipulation.

A training-free framework for efficient visual token pruning in vision-language models.

A wavelet-based inductive embedding method for logical query answering on large Knowledge Graphs.

A framework for iterative tool optimization allowing agents to create higher-order reusable tools from execution trajectories.

A production-assessed benchmark for evaluating the reasoning trajectories of interactive code agents.

Agentic framework for autonomous first-principles materials computation and VASP calculations.

Implementation of GCA for spectral graph denoising and diffusion.

Implementation of Robust Mixture of Low-Rank Experts for adversarial robustness.