Products Live Demo About Pricing Developers Contact
Sign in Get Started
01
Research

Evaluating mathematical reasoning beyond benchmark leakage

Why standard math benchmarks overstate model ability, and how we built an evaluation suite that doesn't leak into training data.

June 2026
02
Engineering

How SRRT Code reads a ten-million-line monorepo

The retrieval and indexing pipeline behind SRRT Code's ability to reason across huge, unfamiliar codebases.

May 2026
03
Product

Inside the SRRT Quant backtesting engine

A look at how SRRT Quant simulates strategies against historical market microstructure data.

May 2026
04
Research

Citation-grounded summarization in SRRT Research

How we constrain SRRT Research to only make claims it can point back to a specific passage for.

April 2026
05
Mathematics

Teaching SRRT Math to show its work

Why step-by-step derivations matter more than final answers, and how we evaluate reasoning traces.

April 2026
06
Company

Why we're building four products instead of one

The case for domain specialists that share a common substrate, instead of one model stretched thin.

March 2026