Sagar Jounkani, Senior Manager – Data Science, shares how curiosity, ownership and a dislike for staying comfortable have shaped his near-five-year journey at Swiggy.
Some people measure career growth by titles. Sagar Jounkani is someone who measures it by discomfort.
The kind that tells you the problem in front of you is worth solving, that the people around you are sharper than you expected, and that you have not yet figured out everything there is to figure out. That feeling is what brought him to Swiggy, and if you ask him, it is what has kept him here.
He joined in December 2021 as a Senior Data Scientist, working on problems most people never think about. How do you predict the exact distance a delivery partner will travel before a trip even begins? How do you turn a vague, typed address into a precise coordinate when the GPS itself has errors?
Today, he leads a data science team and works on challenges that are bigger, messier, and more interconnected than anything he tackled in those early months, but the thing that has not changed is his north star. Find the most complex problem available, and own it completely.
1. Take us back to the beginning. What made you join Swiggy, and what were you working on initially?
I have always been drawn to complex problems that I can solve with data. Before Swiggy, I was at a startup working on last-mile supply chain optimisation. It was interesting, but the work leaned heavily on core optimisation, and I wanted to get deeper into machine learning. I had done my master’s in it and wanted to actually apply it at scale.
Swiggy had both. Machine learning, optimisation, and the kind of pan-India scale where the problems actually matter. That combination brought me here.
I joined the Location Intelligence team as an individual contributor and started with distance prediction for driver payments. When Swiggy shows a delivery partner their payment upfront before a trip begins, that number does not change once the trip starts. So the prediction has to be right before a single kilometre is covered. It accounts for the driver’s current location, likely routes, travel behaviour, and a lot more. I spent the first year and a half on that, and also on geocoding, which is converting a customer’s typed address into an accurate latitude-longitude coordinate, correcting for GPS errors using the textual context they provide.
So essentially, it was AI before AI tools existed to help you build AI. All of it was written by hand, from scratch. Those were the early days.
2. How has AI changed the way the data science teams work now?
Execution speed has accelerated dramatically. Tasks that used to be coded entirely by hand are now largely handled by AI tools. I would estimate that around 95 per cent of the code we produce today is AI-generated. When I joined, it was closer to 5 per cent.
But speed brings its own challenge. When you can build things faster, the new priority becomes quality and intent. You have to make sure you are building the right thing and that the standard does not slip just because the pace has gone up. The problems have not gone away. They have changed shape.
3. A lot has changed since 2021. What feels most different today?
Scale, definitely. We are operating at roughly double what we were when I joined, across both food and Instamart.
But more than scale, the nature of delivery has transformed. Moving from systems built for 30- to 50-minute deliveries to solving for under 10 minutes creates a completely different set of demands. When you commit to 10-minute delivery, the ripple hits everything. Your recommendation systems need to be sharper. Item availability has to be tighter. Average order value has to go up to make the economics work. Everything downstream becomes more complex and more personalised at the same time.
The challenges keep compounding. But that is truly what makes the work interesting.
4. Is there a project that stands out as your most meaningful?
The simplification project I have been leading for the past year and a half. It started from a real, tricky problem.
After a major reorganisation that brought together previously siloed teams working on search, ads, and recommendations, all their systems merged into a single codebase. Each team had built their own systems independently, and suddenly, it all sat together in one tangled structure. We owned the storefront of the Swiggy food app, and iterating on it had become difficult. Changes that should have taken days were taking much longer, and every new feature risked getting lost in the complexity that already existed.
The simplification project was about cutting through that. I have been leading the simplification of the food app’s personalisation models so the Data Science team can be more agile and have higher leverage to drive platform-wide impact, rather than changes disappearing into layers of accumulated debt.
“When you hit a wall of complexity, you end up cutting corners just to get through it. Simplicity gives you leverage back.”
Working on this evolved my understanding of Swiggy’s systems and how Data Science models consistently drive significant business impact 24×7. It helped me develop a deeper appreciation for all the people who have meticulously built Swiggy’s tech systems. Understanding why certain decisions were made, what could be removed without losing value, and how to build in a way that would not create tomorrow’s problems has been one of the biggest things I have taken from my time here.
5. You transitioned from an Individual Contributor to a People Manager. What was the biggest shift?
Honestly, I did not realise managing people was a skill I needed to develop. I assumed good technical judgment would carry over. It does not, not automatically.
My current manager, Srini, helped me understand that. He worked alongside me more like a partner than a manager and showed me how to soften my approach, how to look at situations more subjectively, how to create space for a team member’s perspective before jumping to a solution. The non-technical side of leadership was a new territory for me, and learning it changed how I showed up for my team.
6. What has kept you at Swiggy for almost five years?
The people, honestly. The data science culture here means I am constantly surrounded by people I learn from. And every time I start to feel settled, the next challenge arrives before I have time to get comfortable.
That last part matters more than it might sound. I have switched jobs before specifically because I got too comfortable. That restlessness is real for me. When I have felt like I had stopped growing, I have moved. At Swiggy, I have not had that problem once in all these years. The scale keeps growing, the problems keep evolving, and the people around me keep raising the bar.
7. Which Swiggy value resonates with you the most?
For me, it all comes back to ownership. When you truly own a problem, you do not wait for a ticket to appear or for someone to tell you something needs fixing. You see it, you take accountability, and you solve it. Once that mindset is in place, everything else follows. Collaboration, humility, the drive to keep improving. They are all subsets of ownership.
It is also what I tell my team. If you exhibit ownership, everything else gets taken care of. You do not need to worry about the rest.
8.What is one thing you had to unlearn at Swiggy?
The habit of staying in my lane and focusing only on my scope.
When I joined, I drew clear lines around my system and took time only for what was directly in front of me. But that approach quietly limits you. Over time I learned to deeply understand the bigger picture. The team members I see growing fastest here are the ones who do not operate that way. They understand the engineering systems around them, the analytics systems, the product thinking. They become the most valuable people in any project because they can bring things together across disciplines in a way that staying in your lane simply does not allow.
Letting go of that instinct changed how I work, how I lead, and how I think about what good actually looks like.
9. If you could meet the Day One version of yourself, what would you tell him?
Be genuinely curious about the things happening around you, not just the things directly in front of you. Go and talk to the person solving the problem next to yours. Understand why they are solving it the way they are. See how your work connects to theirs. That habit, more than anything else, is what compounds over time.
You will probably never know everything. But if you keep learning and keep taking ownership, you will always find your way forward.
