Behind TechScribr

Making complex systems feel understandable.

Hi! I’m Sayan, a Data Scientist who has spent the last decade turning messy data into meaningful systems, scalable pipelines, and the occasional forecasting model that behaves well.

A decade in data Building useful systems from noisy, real-world datasets.

Production-minded ML From recommendation engines to forecasting at scale.

Clarity first Math, intuition, and practical takeaways without the fog.

The work behind the writing

I’ve built everything from supply-impression forecasting to keyword and audience recommendation engines to agentic AI tools, wrangled 10M-impression datasets, and even convinced Temporal Fusion Transformers to run nicely at production scale. Before all that, I played with geohashes, Spark clusters, and multilingual text models that insisted “Hinglish” isn't a typo.

My journey has taken me through Walmart, Sahaj, and Broadcom, with pit stops involving Kafka storms, PySpark jobs, and deep-dive knowledge sessions because I can't resist explaining things I find cool.

I earned my Master’s degree at IISc Bangalore, diving deep into Bitcoin’s stability and anonymity. That's far from ML, but a great reminder that exploring different domains often leads to unexpected passions.

Why TechScribr exists

This blog is my cozy corner on the internet where I break down ML and deep-learning concepts with clarity, curiosity, and a little personality. Expect math, intuition, practical takeaways, and the occasional plot twist delivered by real-world data.

If that sounds like your kind of learning vibe, welcome aboard.