$ arunya build --for humans

We buildproducts, AI agents, interfaces, AI appspeople love.

Arunya is a technology research & development lab. We research, prototype and engineer AI applications, agents and product-grade software, with a user-first mindset.

agent / vendor-onboardingRunning
  1. Read supplier documents
    Invoices, registration and bank letters
    00:04
  2. Extracted tax and bank details
    Structured fields, with confidence scores
    00:11
  3. Cross-checked against ERP records
    tool · erp.search_vendor
    00:19
  4. Drafting approval summary…
    1 field flagged for review
    00:23
Human review before anything is submitted
An agent workflow we design: tools, guardrails, human sign-off.
// services

From research to running software.

Anything from a two-week experiment to a long-term engineering partnership.

02

AI applications

Search and answers over your own data, copilots and document intelligence, built to be accurate and observable.

RAG · copilots · doc AI

03

Software development

Web platforms, SaaS products and internal tools: clean architecture, fast interfaces, maintainable code.

web · SaaS · internal tools

04

Conversational AI

Chat and voice assistants for support, sales and operations, connected to the tools your team already uses.

chat · voice · messaging

05

R&D and prototyping

Fixed-scope sprints that test an idea with working software, so you decide with evidence instead of slides.

proofs of concept · user tests

06

AI strategy & integration

Find where AI genuinely pays off, choose the right models, and plug them into your stack securely.

audits · roadmaps · integration

$ arunya new --ideaSomething else in mind?If it involves software and a hard problem, we’d like to hear about it.Talk to us
// craft.md

Not AI slop.
Crafted.

AI is in our toolkit, never a shortcut. Every product starts with the people who will use it and is finished to the pixel, so people actually enjoy using it.

  • User-firstResearch, journeys and usability come before features.
  • Pixel-perfectSpacing, motion and every state, checked on real screens.
  • AI with intentUsed where it truly helps, never as filler.
Our mark, placed pixel by pixel.
// how-we-work

A lab mindset with a builder’s discipline.

We start small and learn fast, then put engineering rigour behind what works.

  1. 01

    Explore

    Understand the problem, the people, the data and the constraints. Agree on what success looks like.

  2. 02

    Prototype

    A working proof of concept, tested with real users and real data, not a demo on a happy path.

  3. 03

    Build

    Production-grade engineering: evaluations, security, monitoring and documentation.

  4. 04

    Evolve

    Measure, improve models and prompts, and scale what works. We stay on as long as you need us.

// what you can count on

  • Research before codeAssumptions tested early, so you don’t pay to build the wrong thing.
  • Production, not demosEvaluation, error handling and monitoring are part of the work.
  • Senior and hands-onYou work directly with the people who design and write the software.
  • You own what we buildCode, models, prompts and docs handed over in full. No lock-in.
// about

The person behind the lab

The name comes from aruṇa, the first light of dawn: the early, exploratory hours where new things take shape.

Kanishka Chowdhury, founder of Arunya Technologies
Kanishka ChowdhuryFounder · Lead engineer

A software engineer, tinkerer and product thinker who takes products from first idea to thousands of customers. He has built AI agent builders, agent debugging playgrounds and no-code workflow tools used by enterprise teams.

6+ yrs
building SaaS and AI product interfaces
15k+
workflows created on a no-code builder he engineered
2.5k+
daily sessions on an AI agent playground he built
Connect on LinkedIn

$ arunya contact

Have an idea worth building?

Tell us what you’re working on. We’ll reply within two working days with honest next steps.

Start a conversation