Are tech schools in India preparing students for the AI/ML age?
Premium- AI/ML will likely permeate every field in the years ahead, which means students should not be looking at it purely from a software development perspective.
- An incentive system pulling the best tech talent into software development rather than more cutting-edge areas would mean the country would reach a saturation point soon.
When Rahul Dave, who was teaching data science and AI/ML at Harvard University, went to one of the Indian Institutes of Technology some time back, he was somewhat nonplussed.
He was talking to a group of mining engineering students about exciting new possibilities in using artificial intelligence/machine learning (AI/ML) for oil exploration. His pitch to them was that their value to oil companies would rise steeply if they were to add AI/ML on top of their core engineering discipline. “And they were like, ‘Nah, we’re going to become software developers," says Dave, who was bemused because he had seen in the US how the demand for AI/ML analysts had already spiked in a variety of fields.
Subhendu Panigrahi, co-founder of Skillenza, which helps companies hire tech talent by running online challenges and hackathons, is not so surprised. “Roles like software developers and full-stack engineers are always in demand, and even more so after covid as every organization wants to digitize their workplace. So a lot of young techies in India just want to become full-stack engineers and join a multinational company," he says.
“They’re not sure they can land good jobs so easily in data science. You can also see this in the bootcamps that companies are launching in Bengaluru. They’re mostly for software developers and full-stack engineers and not machine learning or data analytics bootcamps."
Dave feels an incentive system pulling the best tech talent into software development rather than more cutting-edge areas would mean the country would reach a saturation point and never rise beyond that level. But he had a different experience when he ran a data science bootcamp in Bengaluru prior to co-founding Univ.ai, an online platform for teaching AI - currently only for students from India.
“The students who came for the first bootcamp - most of whom were not from the IITs - had a hunger to learn. A number of them later became TAs (teaching assistants) at Univ.ai and they’re super good and actually better than the TAs I had in Harvard," says Dave. “So the talent pool is definitely there, but it’s not being developed in a great fashion. Once I realized this from the bootcamp, I was totally sold on the idea of doing Univ.ai."
Skillenza is also seeing the winds of change blowing. “A job description opening up in a lot of companies is data engineering. So we’re going to build a bootcamp around it, working with the industry to develop the curriculum and find the faculty," says Panigrahi. “Full stack development is the staple, like dal-rice. Then you add sabji (vegetables) to make the meal complete. Data engineering is one area which we think will become huge in the coming days."
Data engineer is the one who writes code to handle gigabytes of data coming from social media posts or video cameras or geolocation tracking of phones or sensors on machines. He formats the data to make it easier for an AI engineer to build analytical models for it. So the two go hand in hand. Bootcamps like that of Skillenza reflect the gaps in tech education in India.
AI/ML, in Dave’s view, will permeate every field in the years ahead, which means students should not be looking at it purely from a software development perspective. One of his students in the data science course at Harvard was an equine biologist. “It turns out that horse ecology requires a lot of statistical slash machine learning skills," says Dave. “I’m trained as an astrophysicist and my own field has been completely revolutionized. That’s actually how I got into machine learning because with the improvement in chips in the early 2000s, we started not being able to keep up with the data we were getting on our telescopes. We had to start doing it automatically."
He feels textbooks written for any field ten years from now will have parts of machine learning in it. “I actually think it’s becoming a fundamental part of literacy needed for all fields. It’s not quite there, and I think our Indian education system will take a while before it catches up."
Despite shortcomings in their math or programming backgrounds, Dave finds the students he’s teaching on Univ.ai improving by leaps and bounds, which suggests that their inherent qualities just hadn’t found expression earlier. He draws an analogy with how the IPL (Indian Premier League) has been pushing out new talent at such a rate that “Ian Chappell commented in his article on Espncricinfo that every other cricketing nation should be afraid."
“My thesis is that there’s similarly a large untapped pool of high quality students (for AI/ML) in India who don’t even know what they’re worth at the moment," says Dave.
Malavika Velayanikal is a Consulting Editor with Mint. She tweets @vmalu
