Their world
Zerodha runs India's largest retail broker on a technology team of about thirty.
That restraint is a choice, and it is the edge. The question the next year asks is how far a deliberately small, open-source-first team can stretch as the product surface keeps widening.
Their pressure
The lean team is carrying more, against a faster field.
- Team Zerodha's technology team has stayed at roughly 30 engineers for about seven years.
- AI in-house Kailash Nadh says AI is already used internally for engineering, debugging, and call-quality analysis.
- On Claude Kite MCP, built with Anthropic, lets users pull portfolio insights through AI assistants like Claude.
Meanwhile Groww has widened its lead in active users. Doing more without doing it with more people is the whole game — and it is the one you have chosen to play.
The bridge
Meridian AI extends a small team without adding to it.
We are a boutique applied-AI studio that embeds with your engineers and ships production-grade AI on your existing stack — retrieval over your own data, agents wired into your current services — in six weeks. We build on what you already run, with human-in-the-loop controls, so your team keeps ownership and there is nothing new to operate.
One proof
We ship on your existing stack — retrieval over your own data, function-calling agents wired into your current services, and evals plus guardrails in CI. No model training, no rip-and-replace, no new infrastructure to run.
One working session
Find the one workflow a small team should stop carrying by hand.
Twenty minutes on one Zerodha workflow — support triage, ops, or internal tooling — mapped to what an agent can carry and what stays with a person.
See the 20-minute teardown for Zerodha