I'm a product manager who ships data and AI products where the output has to be right, not just plausible. Ex-software engineer (Walmart Global Tech), ISB MBA, now running provider-data validation products for Fortune 10 - US health plans at a Series B AI company. This portfolio shows how I frame ambiguous problems, decide with evidence, and measure what shipped.

What I'm known for:

• Data and experiment driven: I find what everyone assumed was true but nobody checked.
• End-to-end ownership, with close attention to detail.
• Customer empathy: customers like working with me because I take their problem seriously.
• System design: I was a senior developer before I moved to product.

🌟 Case Study Library

Untitled

🚀 How I build

📄 Publications

Two peer-reviewed IEEE papers from my engineering years. Both are classifiers where a miss and a false alarm cost different things, the same trade-off I now manage as precision targets in product.

1. Multilingual Obnoxious Message Classification using BERT

Nikhil Banka (first author), S. Narang, M. Gupta, A. Sharma, D. Upadhyay · IEEE IC2E3 2023 · cited 2×

• Problem: toxic comments (threats, insults, obscenity, identity hate) spread faster than people can moderate them.