Karun Thankachan

About

I didn’t grow up dreaming about working in tech. I grew up in India.Engineering was just… the thing you did.Safe. Respectable. Predictable.So I did it. Software Engineering came next.Also safe. Also logical.At least that’s what I told myself.For a while, I was just going through the motions.Writing code. Solving bugs. Managing deployments.Nothing magical. But something slowly changed along the way.Conversations with seniors & peers on tech advances made me curious.I found myself in hackathons, and weekend side projects - building for building sake.The hustle and grind translated to building solutions at work that impacted others.Some of it made it into IEEE conferences, some into patents, & some led to millons of dollars of impact.And somewhere in between, I realized something simple. Work is meaningful when you build stuff people find useful.Not titles. Not salary. Not fancy tech. Just solving problems to build useful products.Along the way, I also got exposed to great educators. Great ideas. Great engineers who think very differently.And I started enjoying something else unexpected.Explaining things.Not just building them, but breaking them down so others could see what I was seeing.I loved seeing that "aha" moment when the concept clicks for another person.That became another obsession.Today, I’m a Senior Data Scientist at Walmart. I work on large-scale applied ML systems, recommender systems, and agentic AI. I’ve:• Built Applied ML systems across Amazon, Dell, Walmart that generated $100M+ impact• Filed ML patents and published research at top conferences like ICLR, AAAI, ACL etc.• And ended up speaking at places like CMU, Rutgers, USC, Google, Microsoft etc.But none of the above is what's the most interesting thing. The most interesting thing is what's coming next in tech.The age of LLMs is leading to • Shorter development cycle for Applied ML models• Weekly advances in ML and LLM research• New product paradigms in almost every businessAnd that’s what I write about here.• How to build applied ML and Agentic AI at FAANG production-scale• Breaking down recent research advances in ML/DL/LLMs and what it means for industry folk• And how to survive in the industry without getting replaced by AI