Skip to content

KRITRIMAM RESEARCH LAB — KERALA, INDIA

Most AI fails quietly, in the gap between what a system knows and what it confidently says. We build in that gap — retrieval, grounding, evaluation. Two things so far: NEST.ai, a co-tutor that studies your own material. Budhi, AI built into how a business already works.

Explore NEST.ai
Work with Budhi
WHAT WE'RE BUILDING

  • GROUNDED BY DEFAULT

    Every answer traces back to a source you gave it.

  • BUILT FOR THE LONG FORM

    Retrieval and evaluation, not prompt tricks.

  • KERALA, SELF-FUNDED

    Small team, early, no hurry to be acquired.

WHERE TO START

Two ways in.

The same research underneath, pointed at two very different problems.

FOR STUDENTS & EDUCATORS

NEST.ai

A co-tutor that reads what you're actually studying — lectures, PDFs, slides, handwritten notes — and answers from that, not from the internet's average guess.

Coming soon — join the list

FOR BUSINESSES

Budhi

AI agents and retrieval systems designed around your data and your workflow, not bolted on top of it.

Talk to us

COMING SOON

NEST.ai

An AI co-tutor that works from your own course material — lecture videos, PDFs, slides, notes — instead of guessing at the internet's average answer.

Line-art brain wrapped in orbital rings

How it works

  • 1

    Give it your material

    Lecture recordings, PDFs, slide decks, handwritten notes. Whatever the course actually runs on.

  • 2

    Ask it anything

    It answers from your sources, and shows you which one. Not a search result dressed up as a tutor.

    bio_lecture_04.mp4 · 12:31
    chapter_9.pdf · p.203
    notes_scan.jpg
  • 3

    Study from it

    Summarize a lecture. Turn a chapter into flashcards. Generate a quiz. Lay out a plan for what's coming.

For educators

Turn a lecture into structured material in one pass — and see where a class is actually struggling before the exam does.

What to expect

Every answer is retrieved before it's generated, and grounded in the source you gave it. That's what keeps it from making things up. It works across languages, not just English. It's built to be something you come back to in week nine, not a demo you try once and forget.

Why we're building it

A private tutor has always been something only some students could afford. NEST.ai is our attempt at closing that gap — the same quality of one-on-one help, available to any student with a phone, not just the ones who can pay for it.

See the project

Two ways in, one lab underneath.Join the list or talk to us directly.

Work with Budhi

ENTERPRISE

Budhi — AI, built into how you already work

Budhi is Kritrimam's enterprise arm. We work directly with businesses to design, build and adapt AI agents and AI-driven features around your actual data and workflow — not a bolted-on chatbot.

Stippled sculptural figure in white points on black

  • Custom integration

    Bringing AI into the tools your team already opens every morning, instead of asking them to adopt something new.

  • Model adaptation

    Grounding models in your own data, so answers reflect your business rather than the general internet.

  • Applied agents

    Agent and retrieval systems built for one specific operational problem, not a generic demo.

  • Evaluation

    Ongoing testing, so the system is still reliable in month six — not just on the day it shipped.

How an engagement starts

  1. 1

    A conversation. What's the actual bottleneck, and is AI even the right tool for it? We'll say if it isn't.

  2. 2

    A scoped build. One problem, clearly defined, with a way to measure whether it worked.

  3. 3

    Then we keep testing it. Systems drift. Evaluation isn't a phase, it's the maintenance.

Vision

Kritrimam isn't set up to ship one product and call it a company. NEST.ai and Budhi are what we've built with what we know so far — proof of work, not the whole point. The point is the lab underneath both of them.

Most AI products fail quietly, in the gap between what a system actually knows and what it confidently says. We care about that gap more than almost anything else — about where an answer comes from, whether it's grounded in something real, and whether a system can tell the difference between knowing and guessing. That's what we mean by data clarity: not more data, clearer footing under it.

The difference between knowing and guessing is the whole problem.

A glossy black figure lit by a single band of prismatic colour
Colour, used once — the difference between knowing and guessing.

We're building toward the kind of research work usually reserved for a handful of labs in the world. That's a long way from where we are today — a small team in Kerala, self-funded, early. But the foundation we're laying now, in how we think about retrieval, evaluation, and what it actually means for a model to understand something, is built for that horizon, not for a quick exit.

കൃത്രിമം · KRITRIMAM · "ARTIFICIAL"

EARLY ACCESS

Know when NEST.ai opens up.

One email when there's something real to try. Nothing else.

Contact

If you're building something and think we could help, get in touch.