I'm working at the intersection of bleeding-edge research in LLM behaviour modelling and productizing that for useful applications for work such as automated research, autonomous web browsing, and long-running tool use for complex instructions.

Currently building capable autonomous AI agents for knowledge work with long-term auto-updating memory, autonomous learning, tool use, and self-correction.

I've implemented research papers such as ReAct, Toolformer, Reflexion, Generative Agents, and Voyager as well as several multi-agent frameworks over the last 4 years.

Previously I've also spent time doing growth, sales, and outreach.

➡ Message me right here on LinkedIn

Soham Sarkar

CEO @Scoutflo - Self-service Kubernetes Platform | Open Source | Top 30 under 30 SDR | Ex-Zoominfo

Building Scoutflo, an AI SRE platform that helps engineering teams understand and resolve production incidents faster.

Previously, I’ve spent 8+ years building and scaling SaaS GTM engines across Martech and DevTools. I’ve led 0 to 1 sales motions, closed $4M+ in SaaS revenue, and built repeatable pipelines at early and growth-stage companies.

Kalpesh Bhalekar

Everyone wants to join a startup. Until they actually do.

If you're an engineer, you're not just writing code.
You're fixing customer issues, chasing old bugs, shipping new features, managing tech debt, and deploying at 2 a.m. while customers wait for answers.

If you're a PM, every choice you make influences how long the company survives. Three wrong feature bets and your runway might vanish.

If you're in sales, every deal is a make-or-break moment.
You start with no playbook, no tested pitch, no confident price point. You learn everything by doing, failing, and trying again.

Startups are brutal.
But that same chaos is what makes them special.

Our founding engineer, Satyam Pariyar was a commerce student from a small town in India.
Today, he decides how our AI is built, every feature and every release.

Atharva Bondre started in sales, moved into customer success, and now drives what we build next as a PM.

You don't just grow in a startup. You compress years of experience into months.
You learn to take ownership, solve problems end-to-end, and do more with less.

And then there's equity.
Real skin in the game. The kind that can turn two years of work into something that changes your life, maybe even your generation's.

Startups aren't for everyone.
But if you survive the chaos, you'll never want to work any other way.

PS: We're hiring, DM if you're excited to build AI SRE at Scoutflo

Building AI SRE | Co-Founder at Scoutflo | Ex-Zoominfo, Ex-Highlight (YC W23)

I know this is a late New Year post, but 2025 taught me a lot as a founder.

We moved into a new office.
We hired people who genuinely raised the bar.
We made a hard decision to pivot. Not an obvious one, and definitely not an easy one.

Most importantly, we built something real.

We chose to work on one of the hardest problems in engineering: 'Production reliability'.

We ended up building an AI SRE that:
 • Understands a company’s system and operational context
 • Helps resolve production alerts real time
 • Automatically generates post-incident reports
 • Automates day-to-day operational tasks

2026 already feels exciting.

This January, we’re going live with the public launch of Scoutflo AI SRE. What excites me most is that we’re finally at a stage where the product speaks for itself.

Grateful to our customers and our investors! 

If you’re curious to try Scoutflo AI SRE before the public launch, feel free to DM me. Happy to share beta access.




Sanjay Mehta Vatsal Kanakiya Ninad Karpe Arjun Pillai Prasanna Venkatesan
Vedant Vyawahare Atharva Bondre Satyam Pariyar Deepak Suthar

I thought food delivery wasn't real innovation.

I wrote it off as convenience dressed up as tech. Then two experiences over the last couple of weeks changed my mind.

First Incident

I ordered lunch on Zomato. A few minutes later, I received an IVR call.

It said my order was being delivered by a delivery partner who is a person with a disability. He used specialized bike equipment and wouldn't be able to come up to the office.

I was given a choice.
Press 1 to continue with the same delivery partner and come down to collect the order myself.
Press 2 to have the delivery reassigned.

I pressed 1 and went downstairs. That moment stayed with me.

I met a delivery partner who couldn’t walk and operated a custom vehicle entirely for delivery.

Someone who might otherwise be excluded from the workforce was earning independently, supporting his family, and doing so with dignity.

It made me realize what platforms like this can unlock. Not just income, but stability. A shift in how disability is perceived.

Incident 2

One of our employees fell sick in the middle of the workday. They lived far away and hadn't brought their medication. We didn't know a nearby pharmacy.

So we opened Zepto and ordered from the pharmacy section.

Some medicines required a medical prescription.

Within a minute, we received a call.
A doctor was on the line.

They asked about symptoms, age, and medical history. We uploaded the required details. Everything was verified properly.

Minutes later, the order was approved. In under ten minutes, the medicine was delivered.

That was experience designed with safeguards to solve a real problem when timing actually mattered.

What changed for me, both of these experiences are deeply rooted in the Indian context.

They show that innovation isn't always about flashy features or new buzzwords. Sometimes it's about designing systems that work for real constraints, real people, and real moments.

This ecosystem isn't perfect. It still needs accountability and better worker protections. But it is evolving in ways that deserve recognition.

So credit where it's due.

To Deepinder Goyal Albinder Dhindsa Aadit Palicha Kaivalya V. Sriharsha M. Amitesh Jha for building systems that quietly improve lives.

Sometimes innovation doesn't look revolutionary.

Sometimes it just shows up when you need it most.

Startups are hard. Especially in the early days.

Behind the scenes, a very small team is trying to prove that an idea actually holds value, often with limited resources, limited time, and only a few real shots at getting it right. Every startup gets just a handful of darts. You hope at least one hits the bullseye.

Over the last few years building Scoutflo, we've pivoted.
We've had things not work out.
A lot failed before things started making sense.

But today, we're at a place of real clarity.

We know why we're building what we're building.

AI for SRE is clearly needed right now. Every engineer we speak to immediately resonates with the problem. When something breaks in production, the first questions are always the same:
Why did this happen?
How do I fix it fast?

That's exactly what Scoutflo is working towards.

I want to be open about this moment because it matters to us. 

We've built a product we're genuinely proud of. It stands up in conversations everytime during a demo.

We're gearing up to launch on Product Hunt soon, and this feels like one of the most meaningful milestones in our journey so far.

If you've supported us in any way until now, thank you. Truly.
And if you're open to supporting us during this launch, it would mean a lot. A share, a comment, or a DM goes a long way.

If you're interested, reach out. I'd love to share how you can help and what we can do in return.

Grateful for the journey so far and excited about what coming up next.


Vedant Vyawahare Atharva Bondre Satyam Pariyar Deepak Suthar Vatsal Kanakiya Sanjay Mehta Ninad Karpe Arjun Pillai Prasanna Venkatesan

I wrote a complete guide to use Lovable for slide decks.

I built this entire deck in 5 minutes: https://lnkd.in/d8hrR26c

Here's what the guide covers:
• Designing decks as interactive systems, not static slides
• Prompting and navigation patterns that reduce rebuilds and rework
• Achieving modern UI, motion, and polish without burning credits or time

Comment: SCOUTFLO to get the playbook on building decks with Lovable.

Real SRE learning doesn’t happen in courses or docs.

It happens in production.

Most SREs and DevOps engineers are learning in silos.

The only real way you become good at DevOps is by working on live systems.
Real incidents. 
Real failures.
Real pressure.

If you haven’t lived through production breaking at 2 AM, your learning curve is fundamentally different.
And that’s where we saw a gap.

A lot of companies maintain SRE playbooks and best practices internally. Sometimes they’re not even shared across teams. Often, people don’t even know the right way to approach repetitive production issues.

So we decided to change that. We’re open-sourcing SRE Playbooks by Scoutflo.

Completely open source.
Anyone can contribute.
Anyone can learn.
Anyone can make it better.

This is just the start of what we want to do for the community this year.

If you’ve built playbooks, learned something the hard way, or solved a production issue you wish someone had told you about earlier, this repository is for you.

Here’s the link. : https://lnkd.in/egNnw5Ma

Feel free to add your playbooks. 

Our team will be closely working with the community to grow this together.
If you want to help build this community, comment below or DM me.

taplio

The Battle

Soham Sarkar

Kalpesh Bhalekar

Followers

Posts

Engagements

Likes

Performance ScoreKalpesh Bhalekar's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Soham Sarkar

Kalpesh Bhalekar

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Kalpesh Bhalekar's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreKalpesh Bhalekar's LinkedIn Performance