How we cut a healthcare partner's LLM spend to $0 per chart
INSIGHTS

What we have learned building practical, private AI for healthcare.

Short, opinionated notes from our own projects and demos. Every topic has our verdict, the evidence behind it, and a way to try it.

How to read a topic

Each topic carries one of four stances, like the rings on a technology radar.

Use
We use this on real projects and recommend it where it fits.
Try
Promising and worth a trial, but measure it on your own data first.
Watch
Interesting, with open questions. Keep an eye on it.
Avoid
We have had poor results, or the risks outweigh the benefit.

All topics

4 topics

Use Fine-tuning

NER: general vs fine-tuned, and when to switch

Start with a general NER model to learn what you need. Fine-tune a small model once your entities are specific to your domain, and handle negation and context separately.

Try RAG

RAG reranking: when a second pass pays off

Reranking is worth trying when your retrieval is close but not reliable. It is not free accuracy, and the only way to know is to measure it on your own questions.

Try LLMs

Small LLMs on-device: what a 0.5B model can do

A small local language model is a good fit for narrow, text-in text-out jobs where the answer is already in the input. It is a poor source of knowledge, and it needs a person checking its output.

Use Fine-tuning

Small fine-tuned models vs per-chart LLM calls

For a high-volume, privacy-sensitive coding workflow, a pipeline of small, task-specific models beat repeated LLM calls on cost, speed, privacy and recall. We would not assume that holds for every task.

READY TO BUILD WHAT'S NEXT?

Not sure where to start? Start with the problem.

Want this kind of thinking applied to your data? Tell us the problem and we will tell you what we would do.

View All
Contact us

Have a problem worth getting right?

Send us the paragraph you'd normally write to a colleague. We'll get back to you within 24 hours.

hello@tathastha.com