https://youtu.be/0k6OrCkAEdc

Transcript 

Darshan Doshi (00:08)

Welcome to Dasar. I'm Darshan and we have an awesome podcast today. I have with me my close friend Rohit Pandharkar. A quick introduction of Rohit, Rohit is the Global Head of Data Science at OLX Autos. Previously he was the Head of Data Science at Mahindra Group and then also involved with one of the fastest growing startups out of India and the US. He's an MIT Media Lab alumni and has been working in the artificial intelligence, data science, machine learning space for a very long time. He has huge number of use cases where he has helped solve problems, business problems, life problems, customer problems through the use of this technology. So we have with us someone who is in the trenches working, building data science teams, solving critical problems, but at the same time has an exposure to various technologies. And in this podcast we're going to cover a lot of topics around AI. What is AI? What is the difference between AI and Analytics? How do you hire data science teams? What kind of representation is needed of AI within large corporations in the country or around the world? We're going to talk about how AI can improve human performance and make us super humans, if I may call them. Lastly, we also cover the topic of Singularity and we'll get insights from Rohit on how will AI converge into Singularity. So stick in. This is a very long podcast. It is a fast moving podcast. I hope you like it. If you like it, please share it in your network. So let's get started. Welcome Rohit. I'm so excited to have you here. I've given a quick overview of who you are, but let's get straight into it. Most startups, most people are throwing words around like artificial intelligence, right? They have cooked that into their technology, into their product. But what is the difference between artificial intelligence and basic analytics that you could do in an Excel sheet?

Rohit Pandharkar (02:25)

Yes, that's probably the most thrown around world these days. If you hear any startup pitch, people will say, oh, we are powering this using AI. Oh, this is backed by AI, because of AI this will be better, because of AI this will be cheaper, faster, etc. But not everything in analytics is AI. The way I would like to think about this framework and I often talk about this wherever I speak, is four stages of data science or analytics. First is descriptive analytics, which is just describe what happened, let's say Mumbai rainfall. There was a cloud burst in Mumbai on a certain day. How many mm of rainfall happened? That's descriptive analytics. Second stage is diagnostic analytics, which is why did it happen? Is it some cloud burst? Is it some other climate change that could describe the diagnostic piece of it? Again, this is not really AI. It's again, an analytical way of looking at things. Third piece is predictive analytics, which means tomorrow what would happen, when will it happen, and what would happen? That piece is the real machine learning piece because you are using historical data or some learning to predict what will happen in the future. And fourth piece going beyond that is prescriptive analytics, which is if you are predicting something will happen, maybe something good or bad, how do you make it better? How do you make it happen earlier? Or how do you avoid that from happening? That is prescriptive analytics and predictive and prescriptive analytics piece is actually AI. Whenever you are using historical data or some learning data to predict certain outcomes or prescribe what to do better is actually AI and not just analyzing things in an Excel sheet, but actually using algorithms, sophisticated algorithms like Xg boost, deep learning, etc. is the real AI.

Darshan Doshi (04:17)

Brilliant. One of the things that Dasar we try to do is to simplify, make it really easy to understand, leave out the jargons and focus on action, performance, we talk about productivity, we talk a