hi, i'm sahil :)
I'm the chief of staff at Samora AI. We're building the middle layer between the models and the people who use them in the age of voice AI, the tools and scaffolding that let voice agents actually work in production. My honest title is closer to crisis manager, what that means day to day is that I get to do a lot of different things, and that's the part I like.
I scale revenue: cold calls, cold emails, sales meetings, GTM workflows. I work on customer experience: onboarding, and a fair amount of sitting with clients as a consultant when something breaks. I keep teams aligned with accountability routines, outcome tracking, and a steady stream of experiments. And I keep them going with what is, objectively, a legendary playlist of videos (speeches, movie edits, etc.)
who am i?+
I work as a generalist. The part of the job I like most is taking a principle from one place and using it somewhere it has no business being. I care about results, and about staying close enough to the work to see how my own choices shaped how it turned out.
The best moment of most days is when someone walks up with a problem and asks me to help find the way through it.
how i got here
My first real job was at AIESEC, a student-run organisation that places international students into internships with Indian companies. Then Mentormind, where we sold a program called menternships to Upgrad and a few other education partners. Then I met Kartik Sawhney, who'd studied CS at Stanford and worked at Microsoft, and was starting Samora. I came on as his first hire about a year and a half ago. Today I lead our full-stack recruiting product and run client operations on the platform. I've closed around 15 accounts, manage about 30, and I'm responsible for close to $200,000 in revenue a year. The operations team under me is 12 people part-time, plus 3 full-time prompt engineers I hired and lead.
a few honest things
I prefer small teams and small circles. I'm socially awkward, but I go network and build relationships anyway. I'm not a techie, but I know enough about how systems work, and I love that I can now vibe code my way out of an internal problem instead of waiting on anyone.
three things i'm actually really good at
Running many experiments at once. I like coming up with a slightly mad hypothesis and testing it before someone talks me out of it. Some of our best moments at Samora came from exactly this, like a D2C demo agent we customised per company, pointed at their own execs, and used to sell them their own product, only mentioning Samora at the very end.
Shipping. Prototypes, agent scaffolding, customer dashboards, a deck, whatever the week needs.
Holding a lot at once. One of my favourite memories is running a conference at 21 with my phone ringing close to a hundred times in a day. Multiple projects, multiple stakeholders, keeping everyone in the loop, that's where I love being.
ambition+
My ambition is to become one of the best startup operators of my generation. That means finding the hardest problems available, the ones where the outcome genuinely matters, and owning them from start to finish.
I want to keep walking into rooms where I feel like I do not know enough. That feeling usually means I am in the right room. It is where I learn fastest and where I get to prove what I can do.
What I want most is to do this next to a founder and a team who care about the work as much as I do, and who are building something meant to last. I want to give the next few years to exactly that — working with people like that and compounding what I can do.
operating principles+
These are the rules I actually work by. They come from working inside startups, not from reading about them.
1. Never leave a problem unattended. If I see something that needs solving or could be better, I do not let it go. It may not be the top priority this week, but I make sure it gets picked up.
2. The number of problems you solve matters. I used to obsess over the last one percent. In a startup that is a mistake. There is always another leak in the bucket. The goal is not a perfect patch on one hole — it is getting as much water in and keeping it there. Fix it well enough, then move to the next one.
3. Own the outcome, not the effort. It is easy to fall in love with how hard you worked. I try not to. I hold myself to real numbers and work from them. Do not fall in love with your own medicine.
4. Lead the charge. If I need my team to pull an all-nighter, I am there pulling it with them. And sometimes I will tell someone to take the break while I take the shift. You lead from the front or you do not lead.
5. Low ego. As a generalist I am rarely the domain expert in the room. When an engineer tells me how something works, I listen and adjust. Knowing what you do not know is not a weakness — for a generalist it is the whole job.
6. Overcommunicate, and be specific. Be specific about what you are solving and why. Then say it again. In a fast-moving team things rarely land the first time, and a vague message is worse than no message.
case studies+
AIESEC
Context: AIESEC is a student-run organization that places international students into internships with Indian companies. Ran the India chapter for a year, out of Bombay, at age 22. The job was really category management.
Problem: Long list of internship opportunities, most unserviceable — had to find applicants for live openings fast and sort which were worth carrying, and teach all of this to college students running city teams across the country.
Solution: Started from first principles. Took every application received, pre and post pandemic, and broke it down by job role, country of origin, and country the applicant wanted to go to. Created targeting guides for which roles and countries were worth pursuing. Built an AI-plus-human screening process.
What I'd do differently: Would have listened to the data more carefully. The patterns in how people behaved were hard to beat without a real structural shift in the world, and we were too optimistic to listen.
Samora
Context: Early Samora, before YC and before any real credibility — a handful of people selling a voice AI platform.
Problem: Two problems — selling a product that barely existed with no brand, cold-messaging VPs and PMs at large companies; and what was being promised ran ahead of what engineering had built, with every client wanting something custom.
Solution: For the first, studied how other agentic products sold themselves, taking inspiration from Relevance AI's signup flow, and made demos hyper-personalized from the client's own URL and public data. For the second, made client KPIs the product KPIs, and over time several SOPs turned into product features.
Outcomes: Those demos brought in The Hiring Company and Successbooster — both still the largest accounts, contributing about $5,000 MRR each.
What I'd do differently: Would have made us less reactive, sooner. Would have started instrumenting product usage and analysing it earlier to lead the market instead of following it.
the work, in four lanes+
— sahil :)