active

Cortex

The simplest semantic layer for data driven applications.

2015

i. support · 2015–2018

Three years watching analytics fail everyone it was meant to serve.

Five analysts, fifty teams, one question phrased five ways. Nobody could self-serve. I learned SQL just to do my own job.

500+ conversations. All resolved too slowly.
The complaints became a spec sheet.

Queue: 0

SLA

78%

2021

ARR Growth

+12%

Q1 → Q2 → Q3

Healthy?HEALTHY

ii. founder · 2021–2024

Built a customer success product from scratch. Talked to 300 operators. Killed it.

One engineer, one market. Backend, frontend, infra, integrations. $125K pre-seed on a SAFE.

Five tools. Five definitions of "healthy." Zero shared language.
The real problem was underneath.

NOW

iii. the pivot · 2024–today

Not a CS problem. A foundation problem.

Every interview pointed the same direction: the gap between "we have data" and "we understand it" is structural. Better UIs on broken foundations deliver confident wrong answers faster.

Text2SQL gave confident wrong answers. Worse than nothing.
Killed the product. Kept the problem.

v1 · CS tool

Telescope for CS

Wrong layer.

v2 · now

Cortex

New foundation.

Not another layer on top. A rethink of what's underneath.

400+ tools. same 5% served.

Data analysis should be for everyone.
Not just for those who can afford a data team. Not just for those who already know what to measure.
Read More

Epiphany

Peel all the layers off the onion and solve the problem at it's core

Status

Open source · MIT

Stack

Python

FastAPI

Vue

TypeScript

PostgreSQL

Redis

Docker

Say Hi!

Builders, backers, design nerds, future co-conspirators. Email's fastest, but I'm everywhere below.

GitHub

@shankster

in/nilotpal-roychowdhury

Twitter

@init5n

nilotpal@jointelescope.com