New Year
After dining in at The Imperial, Koramangala, me and my co-founders welcomed 2025, watched sky shots fire from Tamal’s(our friend) rooftop. We received a few orders that day on Craveo app. We were facing high churn rates and the numbers weren’t exciting. Our onboarding funnel had friction.
We spent most of Jan & Feb optimizing the onboarding funnel because that was the USP of Craveo. Shashank and I spent time to pull more users in the app through our Instagram and reddit campaigns.
We revisited, brainstormed around possibilities of price comparison applications surviving largely duopoly market 10 years from today.
With SkyScanner, the flight booking market is overly saturated and comparatively, it is massive.
Building in the consumer space requires high cap-ex. We operated in a budget category and it was difficult to monetise the model.
Remy
While iterating and studying the bloating AI trend, we concluded with a pivot. We moved to ‘AI powered Food Ordering’, we called it Remy. The problem statement: Users spend 30 to 80 minutes deciding what-to-order and where-to-order from. With Remy, users could place their food orders in just 2 minutes.
We received a shoutout from Caleb, which received decent engagement on X and Instagram. We received our first 1K initial users from Caleb’s post. That helped us test the product with real use cases. We found issues, we fixed them. We received few investor engagement, but were rejected with most feedbacks being “you guys are still early”.
We grew upto 10k user base in a 2 months, but retention was low. We shrunk the use case to late night and budget users. Remy was very good at finding budget biryanis under ₹200 at 2 am. We experimented and kept iterating, but failed to improve the numbers.
By end of August, we shut Remy.
LLMLab
At many times, to pause is the right decision. True builders never stop building.
2025 was a very important year. People were engineering great products in SF, in HSR & in rest of the world. We ran short on our runway.
We brainstormed different businesses, problem statements and markets that shall stay 10 years from today. Although crowded, GEO is an emerging market. With Google pushing it’s AI Overview on 16% of it’s searches and ChatGPT with it’s massive distribution so early, this must survive.
We talked to a few companies, convinced two of them to become our design partner. Instead of building the product, we went manual(‘Do things that do not scale’) to study the fundamentals. How does search work under that minimal chat ui.
Drafting this on 31st December’25, we have converted two of our design partners into paid customers. We have also onboarded an enterprise.
2026 excites me already to build, to learn and to execute.
Best friend’s wedding
We have shared classroom for 12 school years and are best friends ever since. It was a weird feeling, yet an amazing experience.
