Mathangi Sri

Mathangi Sri

Co-founder

About

Mathangi Sri is the Co-founder of YuVerse, where she helps enterprises unlock efficiency and elevate customer experience through AI. YuVerse's solutions — spanning alternate-data scoring, intelligent document processing, and AI-powered voice automation — enable organizations to accelerate credit decisions, sharpen fraud detection, and automate critical workflows from credit report generation to collections.

Before founding YuVerse, Mathangi served as Chief Data Officer at Yubi, where she shaped the company's data strategy, governance framework, and AI roadmap. Her experience building high-impact, data-driven systems at Yubi — and across financial services and technology — laid the groundwork for YuVerse.

Mathangi holds over 100 patent grants in intuitive customer experience and user profiling, and is the author of two widely regarded books: Practical Natural Language Processing with Python (Apress/Springer) and Capitalizing Data Science (BPB Publications). She has built and scaled data science organizations at Citibank, HSBC, and GE, as well as high-growth startups including [24]7.ai, PhonePe, and Gojek.

A passionate advocate for applied AI, she contributes actively to the data science community through lectures, writing, advisory roles, and mentorship — and serves as guest faculty at IIIT Sri City, IIM Kashipur, and NIT Trichy.

Her leadership has earned wide recognition, including Trailblazer Visionary in AI (ET Edge, 2024), 10 Most Influential Leaders in BFSI (Analytics India Magazine, 2024), CDO/CTO/CIO of the Year (Bharat Fintech Summit, 2024), and Inspiring Woman in AI (3AI, 2024). She has also been listed among India's Top 50 Influential AI Leaders (Analytics India Magazine, 2021), Top AI Leaders in India (3AI, 2021), and was named one of "The Phenomenal SHE" by the Indian National Bar Association (2019).

 

Agentic AI promises autonomous workflows. The reality is messier. Agents that work in demos break in production because they lack what we call the harness — domain knowledge, process awareness, cultural context, and the ability to handle the exceptions that define real enterprise workflows. This talk shares hard-won lessons from deploying agentic AI at scale: voice agents handling 2.5Cr+ calls a month who need to know what "maafi ho jayegi" means for collections outcomes, document agents that must navigate smudged land records in local languages, and credit decisioning agents that need to match analyst judgment — not just generate text. The argument is simple: LLMs give you intelligence. Harness engineering gives you outcomes. And in enterprise, only outcomes matter.

Read More →