Approximate Labs

Boulder, CO

Independent research into the efficiency of reasoning.

We build compilers, benchmarks, and self-modifying architectures to enable System 2 thinking in Large Language Models. Moving beyond static context to active state management.

Research Log

2025 Agenda

[Redacted] Benchmark

Nov 25, 2025
Evaluation

Upcoming open benchmark, this is a placeholder.

Lab History (2022–2024)

TabLib
Dataset

The world's largest open-source dataset of tabular data. 627M tables extracted for training Large Data Models.

HuggingFace ↗
Sketch
Tool

An AI code-writing assistant for pandas that understands data context via approximate summarization algorithms.

GitHub ↗
Julyp
Product

Data-focused AI assistant (formerly Tabby Chat and Julyp) backed by on-demand Jupyter Lab environments with dynamic installs, cached data pipelines, dashboards, and full iframe/canvas rendering.

TableGen
Product

Experimental smart spreadsheet where each cell runs an agent with row and column context. Agents execute in parallel to fill tables and enable fast data manipulation.

Principal Investigator

Justin Waugh

Focused on the intersection of evolutionary algorithms, compilers, and reasoning efficiency. Previously at Unsupervised and Approximate Labs (v1). Background in Physics (University of Colorado Boulder).

Github @justinwaugh