According to Umang AI is all about transforming raw data into compelling narratives that drive impact.

Can you share the most interesting part of your career journey?

I would choose my first day stepping into the corporate world of data as a defining moment. It was then that I truly realised the power of data, especially when I started working with it hands-on. Since then, I've solved multiple complex problems using data analytics and science, uncovering insightful stories that have helped uplift key business KPIs. It's been all about transforming raw data into compelling narratives that drive impact.

If I have to be very specific, I would say that after completing my master’s, I initially wanted to pursue a PhD. However, I ended up choosing a corporate job instead, and it completely changed my perspective. Beyond working on real-world problems that create an impact, I also got the chance to meet new people and learn from their AI journeys, which has been incredibly rewarding.

What excites you the most about working with data and AI?

One of my seniors used to say, "Where there’s data smoke, there’s business fire." I truly grasped the meaning of this when I began managing a pod independently in an e-commerce organisation. I found invaluable insights that helped product managers shape the product and drive key KPIs for the company. All the hustle was about uncovering a meaningful story behind the numbers. Being a technical-minded individual, I enjoy tackling complex business challenges and strategising how data science solutions can be applied.

Describe an interesting project that you have worked on.

It’s tough. I have been part of many interesting projects and am still working on many more. I'd like to share about one of the projects I had worked on for a B2B organisation. The objective was to automate the mapping of distributors and retailers based on multiple factors, such as GST, DL, PAN, name, address, etc.

It involved extensive data modelling—engaging in long discussions with the operations team, communicating with tech teams and data engineers to understand the multiple microservices and ERP systems, and consolidating all the necessary data in one place.

Once we brought together data from both distributors and retailers, we applied mathematical functions to generate scoring for each match. Then, we used machine learning to train the model and applied a heuristic approach to create three categories: automatic mapping, manual mapping, and discard.

This streamlined process significantly helped the operations team by eliminating irrelevant combinations whilst the automatic mapping improved the retailer experience. As a result, we saw a noticeable increase in revenue.