as an Indian founder, this one hits different.
watching @kunalb11 build CRED from an almost impossible idea into one of India’s most iconic companies, then step into leading WhatsApp globally, is insane.
trust, craft, distribution, ambition.
this is the kind of journey that makes you want to build harder.
The hardest part of building an AI on-call engineer isn't finding the bug. It's teaching it to shut up when it isn't sure. A confident wrong answer in prod is worse than no answer.
A lot of people asked me why I moved to San Francisco before graduation.
The honest answer?
I couldn’t stop thinking about what we were building.
I’d think about it in class.
I’d think about it after class.
I’d think about it while working internships.
I’d think about it before going to sleep.
At some point I had to ask myself:
Am I building this on the side because it’s convenient?
Or am I keeping it on the side because going all in is scary?
I already knew the answer.
So I packed my bags and moved.
Today, four founders live under one roof in San Francisco.
We spend our days talking to engineers, learning how teams handle incidents, building product, shipping updates, and trying to solve a problem we believe matters.
A few months ago, @TasksMind was an idea.
Today we’re working with pilot customers and preparing for closed beta.
Still early.
Still learning.
Still shipping.
Let’s see where this goes.
Happy Friday from San Francisco ☀️
Most people think startup life is investor meetings, launch announcements, and viral AI demos.
Here’s what this week actually looked like:
• Customer discovery calls every day
• Converted a customer into a pilot
• Built and tested new agent workflows
• Shipped product updates
• Got brutally honest feedback from engineers
• Spent hours debating product decisions with the team
• Assembled a mechanical keyboard at Cursor Compile
• Got stuck in downtown SF traffic
• Made homemade pizza with the team
• Touched grass (important)
• Worked from coffee shops, apartments, and anywhere with Wi-Fi
The biggest lesson?
Companies aren’t built during the exciting moments.
They’re built during the repetitive ones.
The customer calls.
The bug fixes.
The late-night discussions.
The constant iteration.
A few months ago, TasksMind was just an idea.
Today we’re working with pilot customers, learning from engineering teams, and getting closer to our closed beta.
Still early.
Still learning.
Still shipping.
Have a great weekend everyone
A few years ago I was figuring out classes in Nebraska.
Today I’m talking to engineering teams in San Francisco about a problem we’re obsessed with solving.
Life changes fast when you keep showing up.
I built the investigation pipeline for our on-call agent, and the hardest part wasn't what I expected.
It wasn't reading the alert. It was teaching it to stop guessing.
Early versions would grab a stack trace and confidently propose a fix for the symptom, not the cause. In production, a confident wrong answer is worse than no answer.
What changed it: correlate the error window against what actually shipped before reasoning about a fix.
Now it shows its evidence and a confidence score, and escalates when it isn't sure.
Most of "debugging" turns out to be connecting signals scattered across five tools. Get the correlation right and the root cause mostly falls out.
VC horror story:
Met with an angel investor who was ex-DeepMind.
I thought I was there to pitch TasksMind.
First, he spent 20 minutes explaining why our startup couldn't work.
Then he spent the next 20 minutes pitching his own startup.
Left the call with no check, no conviction, and somehow a product demo.
Fundraising is a beautiful process.
😭