Newsroom

We're an AI research lab for foundation models. We reverse engineer, interpret and build environments, libraries and frameworks. Our dream is to accelerate ML development with precision and intent for all, enabling safer and better ML across physics, life sciences, industrial, frontier and critical verticals.

Shipping a developer environment for foundation models

Product Now

aq is a developer environment for foundation models built with frameworks, libraries, tools and autonomous agents helping design, test, and refine models. aq is built for precision and intent. Think Next.js for JS, but for ML, handling architectural experiments and building AI the way it should be built, deliberately, not by guessing.

Research Aug 2026

First paper: Finding Usable Weight Mechanisms with Tiled SVD

We published our first paper. We extract mechanism mounts directly from weight sites by column-tiled SVD, each mount a triple as trigger, write, and strength. Pre-registered A/B/C eval on Gemma-2-2B (WikiText-2): 182/182 GO.

Lab Now

Neuro foundation models for disease biomarkers

Mapping correlations between the human brain and LLM neural nets. Building a testbed for drugs and cures aimed at Alzheimer's, dementia, and schizophrenia.

Research Summer 2026

Simulating LLM behaviour across environments

Activation steering for bias, Chess Agents Simulation, Echoes gossip RPG, and Among Us Agents Simulation.

Partners Now

Wound recovery imaging with Zyogen

With Zyogen we are building interpretable medical vision transformers and diffusion models, including training a ViT on wound recovery imaging and interpreting it to distill a prediction system you can still inspect.

Lab Now

Interpreting and quantizing world models

Reading how a model represents the world, then compressing that structure so it stays inspectable and usable.

Lab 2026

$100K in compute and funds

We have raised about $100K in compute and funds through NVIDIA Inception, Emergent Ventures at GMU, Founders Inc, and The Residency.

Partners Now

Continuing with LAION

Building an interpretable VLM that maps text and vision towers on top of OpenCLIP.

Partners Spring 2026

Working with KAIST

Using interpretability tooling for preference detection analysis in LLMs.

Partners Spring 2026

Working with AEIA Lab

Training SAEs on a language model to sense conflict, ambiguity, and tone, and choose the right response.

Product 2025

AQIT: interpretability tooling you can run

Open-source CLI and Python SDK for the gap a loss curve does not explain. Inspect, simulate, watch, diff, steer, probe, and eval.

Research Summer 2026

Experimental weight editor on Pythia 2.8B

Agentic ROME with causal tracing and a validation loop that rolls back and retries on failure. Case studies on factuality, bias correction, and censor auditing.

Research Summer 2026

Ray-Ban glasses social data into training JSONL

How Reddit, X, and Hacker News discussions around Meta Ray-Ban glasses were turned into a structured JSONL training dataset through a multi-stage pipeline.