Sriram Padmanabhan
Products · Systems · Stories
I'm a cross-disciplinary product lead with 10+ years across Software Engineering and Product Management in the US and India. I'm drawn to complex problems that need equal parts user understanding, resilient systems and a story that holds them together.
Right now: Shipping marketplace products at Indeed. Also exploring how AI could help solve previously intractable problems in non-tech domains.
01
Work
-
Hiring marketplaces rarely look past the resume, and when they do they ask the wrong questions. Placing questions where they fit the flow, and choosing them from live employer demand, tripled the structured data available for better matching.
Indeed · Senior PM · 2024
-
Jobseekers apply to jobs they can't get. Understanding job requirements, scoring fit with an ML model, and showing the gaps before candidates applied lifted application-to-hire success 12% in the US.
Indeed · PM · 2021
-
Every Yelp search and business location resolved through one overloaded geocoder inside the monolith. Broke it out into a standalone multi-stage service over unified public and internal map data.
Yelp · Software Engineer · 2018
02
Beyond work
-
Cities worth living in
The gap between what Indian cities are and what they could be is a design, management and funding problem. Good solutions exist, specs are being published, and civic support is often healthy. The challenge: solving it end to end and keeping it solved, at an Indian city's scale and dynamics. The interesting question for AI here is whether it can be built into the administrative layer rather than bolted on from outside. Working with civic NGOs on this.
-
Advising and mentoring
Selectively open to advising founders and early product teams, usually on technically complex domains or two-sided marketplaces. The problems I find most interesting are the tangled ones: where the underlying system, the user's real mental model and second-order effects are all in play.
03
Selected Writing
-
When my news app said the losing party won
My news app reported an election result backwards, and a year-old bill as breaking news. Both were fluent and internally consistent. The lesson is that you don't fix this with more prompt instructions; you stop asking the model to find the news at all.
Read on Medium → -
Living in Nonlinear Times
More people exchanging more information means the state space grows faster than anyone can model it. Predictions break down. The response isn't to forecast harder; it's to build things that stay standing when the forecast is wrong.
Read on Medium → -
It's 7 PM After Work. What Can You Actually Do in Your City?
Cities are measured by what gets built, not by what a resident can actually do at 7pm on a weekday. Products stopped being measured that way decades ago. Cities should borrow the habit.
Read on Medium → -
Home Again: Notes on Life, Change, and Possibility in India
Moved back to India after a decade away. The warmth and the entrepreneurial energy are real. So are the infrastructure gaps, which stop being an abstraction once you are living inside them.
Read on Medium →
04
Let's Connect
Especially interested in hearing from product leaders working on hard AI or marketplace problems, and people building civic or urban infrastructure.