Purva Bangad - Senior Data Scientist and AI Technologist

Purva Bangad is a Senior Data Scientist and AI technologist based in Michigan, with over eight years of experience building applied AI and machine learning systems across large enterprises. She currently serves as Senior Data Scientist at Meijer, where she builds GenAI powered marketing tools on Databricks. Previously, at Fidelity Investments, she architected an AI powered platform using OpenAI and AWS Bedrock’s Claude that automated the migration of ninety thousand legacy business intelligence reports, cutting processing time by ninety-seven percent. Earlier in her career, at Toshiba Global Commerce Solutions, she built a machine learning classification system that reduced manual issue analysis time by eighty percent.

Purva’s path into technology and entrepreneurship began early. She was part of the founding committee of India Billboards, an analytics driven advertising technology startup that was later recognized by SiliconIndia as a Top 10 AdTech Startup. That early experience shaped how she approaches technical work today, treating engineering not just as a job function but as a way to build and ship things that solve real problems for real people.

Outside her core enterprise work, Purva researches the intersection of artificial intelligence, education, and labor markets, with a paper currently under review at IEEE examining how AI can support lifelong learning and career guidance. She is also an active open source contributor and technical writer, publishing on Medium and dev.to about applied AI and large language model engineering. One of her recent projects, an AI powered job application tracker built using Claude’s tool use capability, has been open sourced on GitHub and picked up by other developers solving the same problem in their own job searches.

Purva contributes to the broader technology community as a judge for the Conrad Challenge and the Business Intelligence Group’s Global Awards, and she mentors other data professionals navigating their own career transitions. She holds a Bachelor’s degree in Information Technology from the University of Pune and two Master’s degrees, one in Information Systems and Decision Sciences from California State University, Fullerton, and pursuing Professional MBA.

She is currently working on a AI Student Coach app and early stages of Orchapod AI.

What is your typical day, and how do you make it productive?

My days are simple, but they still feel like they are happening. Work, explore, eat, work out, that is the shape of most days. What makes it productive is what I fill the in between moments with. I listen to podcasts while working out instead of just zoning out, and when I have downtime I like reading about what other people are building and how they are doing well. It keeps things interesting and keeps me learning even when I am not technically working.

How do you bring ideas to life?

I bring ideas to life by trying to implement them, not just talking about them. Sometimes that is the only way to find out if an idea would be meaningful once it exists in the real world instead of just in your head. It takes patience, and it takes being honest with yourself about reality once you start building, because things rarely go the way you imagined when the idea was still just an idea.

What’s one trend that excites you?

Honestly, right now, anything AI is exciting, in a slightly chaotic way. Every week there is something new, and it either teaches you something, keeps you on your toes, or makes you realize you are already behind. I find that mix genuinely fun. It is one of the few fields right now where staying curious is not optional, it is basically required just to keep up.

What is one habit that helps you be productive?

Talking to friends and observing. A lot of my best thinking does not happen at my desk, it happens in conversation or just paying attention to what is going on around me. People say things or do things without realizing how useful that perspective is to someone else who is stuck on a problem.

What advice would you give your younger self?

Just do it. Stop overthinking it. You blink, and suddenly time that felt endless is already gone. I spent too many early years waiting for the right moment or the fully formed plan, when most of the time you just needed to start.

Tell us something you believe almost nobody agrees with you.

AI should make us sharper, not softer. Everyone talks about AI saving us time, but nobody talks about what we quietly stop practicing once a model does it for us. My unpopular opinion is that if AI is not making you noticeably smarter after a year of using it, it is probably making you a little dumber, and most people are too busy enjoying the convenience to notice which one is happening to them.

What is the one thing you repeatedly do and recommend everyone else do?

Experiment. Implement. Take a break. And do not get pulled into social media, because at this point LinkedIn basically is social media too. It is easy to confuse scrolling through what other people are doing with doing something yourself.

When you feel overwhelmed or unfocused, what do you do?

I mostly watch something. A show, a movie, anything that pulls me into a different character, place, or story for a bit. It gives my brain somewhere else to be for a while, and I usually come back more focused than if I had just tried to push through.

What is one strategy that has helped you grow your business or advance in your career?

To love what you do, to respect the people you work with, and to grow with the company rather than only for yourself. That has been true for me from my very first startup all the way to where I am now at a mid sized company. Growth that comes at the expense of the people around you does not really last.

Here is how it actually played out. When you show up caring about the outcome and not just your own piece of it, people start trusting you with bigger problems, not just more of the same work. That is how I started my initial days in the start-up I went from executing someone else’s plan to being handed an open-ended problem and being trusted to architect the whole solution myself. Nobody hands you that kind of ownership because you asked for a promotion. They hand it to you because you already showed up like you had that level of ownership before you officially did.

What is one failure in your career, how did you overcome it, and what lessons did you take away from it?

I have plenty to choose from, but the one that shaped me most quietly was underestimating human opinion. Early in my career, I trusted the numbers to make the case on their own. I built the model, ran the analysis, and assumed a clean result would be persuasive by default. It was not. People did not move because a chart told them to, they moved when someone connected that chart to something they already cared about. I overcame it by learning to pair the analysis with the human story around it, not instead of the numbers, alongside them. The lesson stuck with me. Numbers can do the talking, but it is the human touch that paints the picture and gets it hung in the gallery.

What is one business idea you’re willing to give away to our readers?

Using the data that already exists to help people, instead of letting it sit in a report somewhere. I have been working with IPEDS and BLS data, and one idea I would happily give away is using that kind of data to guide students making education and career decisions, basically gap between completions, degrees and the actual job demand. There is so much advice out there already, most of it generic, and real data specific to your situation would cut through a lot of that noise.

What is one piece of software that helps you be productive? How do you use it?

Visual Studio. It is where I find out if an idea has any life in it. An idea can sound great out loud, but building even a rough version of it in VS Code is usually what tells you the truth about whether it is worth pursuing further.

What is the best $100 you recently spent?

Professionally, it was probably my Claude Annual subscription, though it ends up costing more than $100. But if I am being honest about the best $100, it was not for me at all. I bought my mum a pair of HOKAs. She is in her sixties, fragile but still fit and always moving, and her knees needed the support more than she would ever admit. I had tried and tested them myself first, so I knew they worked before I handed them over. Watching her walk more comfortably was worth a lot more than $100 to me.

Do you have a favorite book or podcast from which you’ve received much value?

I am more of a mystery and thriller reader by nature, so my book answer changes with whatever I am currently pulled into rather than one favorite. On the podcast side, I have been following Raj Shamani. What I get out of it is not any single episode, it is the range. Every episode is a reminder of how much learning and how many different worlds exist outside your own bubble, and how much you do not know that you do not know.

What’s a movie or series you recently enjoyed and why?

Key learnings:

Small, self directed projects built to solve a real personal problem can carry as much career and credibility value as large enterprise initiatives, especially when the process is documented and shared publicly rather than kept private.
Publishing work in progress, rather than waiting for a polished result, creates more opportunities for feedback, visibility, and unexpected use by others than waiting for perfection does.
Technical credibility built inside a single company does not automatically transfer outward. External activities like writing, open sourcing code, and judging competitions are what make expertise visible beyond an immediate employer.
Applying newer AI capabilities, like tool use rather than plain prompting, to narrow and well defined problems tends to produce more reliable results than open ended AI applications.
Overwhelm is often a symptom of too many undefined threads competing for attention rather than genuine task difficulty, and simply writing down the actual problem can clarify what matters most.