The Quanfire blog

We build AI for professional-services firms — including our own domain models where it counts. Here's how we think, what we ship, and the evidence behind it.

Research & Engineering

How we build our own models, and how we measure them.

Research & Engineering

Our first reranker, and the training bug that made it worthless before it worked

We open-sourced rerank-statute-en, our first reranker: a cross-encoder that reorders a retriever's top-k on Indian central statutes. Retrieve-then-rerank Recall@1 goes 0.205 to 0.336 (+63.8%), confidence interval excludes zero. Here is the number, the headroom we did not hide, and the negative-sampling bug that collapsed the first run.

Aug 15, 2026
Research & Engineering

We said our legal model was weak on statutes. Here is the one that is not.

We open-sourced embed-statute-en, a 2.4 MB adapter for retrieving Indian central statutes. On the un-gameable low-overlap slice it more than doubles Recall@1 (+131%), with the numbers, confidence intervals, and provenance to check it yourself.

Aug 14, 2026
Research & Engineering

Retrieval across 16 Indian languages, and the receipts to check it

We open-sourced embed-gov-indic: a 2.4 MB adapter that retrieves Indian government press releases across 16 languages, +27.9% Recall@1 over the base model. The model is public and the method is documented, so you can pull it and judge it on your own text.

Aug 12, 2026
Research & Engineering

Don't take our word for it: read the tokens yourself

Quanfire's multilingual embeddings are live on Hugging Face: Indic-first, openly licensed, trained only on clean, documented data, verifiable in your browser.

Aug 11, 2026
Research & Engineering

The judgment is public domain. The headnote is not.

A court judgment is public domain; the headnote printed above it is not. How we excised the copyrighted layer so a legal embedding ships Apache-2.0.

Aug 10, 2026

For Practitioners

Where AI actually helps professional-services teams — and where it doesn't.