Amazon India launched the third edition of Machine Learning (ML) Summer School on Tuesday. The programme facilitates lessons by scientists at Amazon for students interested in learning about ML technology. The latest edition of the programme will be conducted over four weekends in September.
Due to the strong interest in generative AI, machine learning has seen a lot of buzz recently. Since the pilot of ML summer school in 2021, Amazon said that the programme has attracted a lot of attention. Rajeev Rastogi, Vice President-International Machine Learning, tells Outlook Business in an exclusive interview about the role Amazon is playing in preparing the workforce for rise of AI and how the new technology is shaping customer experience.
Edited Excerpts:
Rising role of artificial intelligence and ML has raised anxieties about the preparedness of the workforce in dealing with new technology. How would you assess the readiness of the industry to deal with the rise of Machine Learning algorithms?
There is an ML revolution that's underway today. I mean, especially with large language models and generative AI, you have the capability now to generate a variety of content, articles, text, blog posts, email, images, video and so on. And ML has been on an upswing over the last one or two decades. At Amazon, we have been using ML extensively across our retail business right from people searching for results or to recommending products. We've been a pioneer in the area of recommendations with item-item collaborative filtering techniques and so on.
ML has wide applications in many different verticals. From manufacturing, financial services, retail and agriculture to health care, you name it. Every industry is embracing ML today and deploying it for a wide range of applications. There is going to be a huge demand for ML talent in the coming years because of all the applications of machine learning. And as a result, I think clearly the workforce in India and around the world will have to be trained, up skilled, and re-skilled for machine learning. We see a huge response to the training programmes that we've been running, like Machine Learning Summer School. So clearly, there's a lot of interest in the area. There are a lot of students who are taking training programmes like ours, but also adopting self-learning based on resources on the web. There are lots of online courses, videos, and tutorial content online.
I'll quote NASSCOM's State of Data Science and AI Skills in India 2023 report, which said that India has an estimated installed talent pool of 416,000 as of August 2022, of professionals with data science and AI skills, while the total demand for this talent pool is around 629,000. So, there's a gap between the total demand and supply of almost 51 per cent. And India is expected to have a total demand of over 1 million professionals by 2026 in this area. But the good news here is also that schools and universities have ML courses, and students are getting trained and learning machine learning.
Making youth future ready would be an important step in the coming years. What role is Amazon playing in skill development?
At Amazon, we believe that success and scale bring broad responsibility. And we have this immense responsibility to nurture and empower future generations. So, we've always worked towards helping young talent gain in demand skill sets like cloud computing. Machine learning is clearly the area that is in-focus today and extremely relevant, and it would help students get into new careers that are in demand and have a lot of opportunities.
ML Summer School is a good example where we are providing students with deep machine learning skills. We are training them through tutorials on machine learning fundamentals, but at the same time, also connecting that with practical applications so that they are industry ready for a career in machine learning. Machine learning summer school combines both theory and practice, so that students get an idea not only of the theoretical concepts, but also how they can be leveraged to solve real world problems in e-commerce and other areas.
We also conduct the ML Challenge, which is our flagship ML competition. It received a total of 25,567 registrations and 10,409 students submitting their solutions from 2,000+ colleges. The one other thing that we've been very particular about, and which is almost a tenet for us, is that we want to democratise machine learning. Our ML challenge is open to all schools and universities across India. It's not just restricted to a subset of schools. And that's why I think the response has been so large in terms of the number of registrations that we get. For example, in 2022, we got 17,500+ registrations for the Machine Learning Summer School. The ML challenge is an ML competition in which we take an Amazon problem and put it out there for students to generate solutions. This is a great way for students to get hands-on experience in solving real world problems. They get real data that we put out there. So, it's really a great experience. And of course, for us, it's a great way of tapping into talent that exists in academia.
After two editions of Machine Learning Summer School, what have been your takeaways? In the coming edition of machine learning school, what changes are you going to introduce given the buzz around ML has increased substantially?
We've had two past editions of the ML summer school. And clearly, one learning and takeaway is that the response has been tremendous from students. So, there's clearly a thirst among students to learn about machine learning and that's super encouraging that so many students actually register for the ML summer school.
In the inaugural programme, we were a little bit circumspect because that was the first time, we were doing it. We had 3500+ students register for the event and we had only advertised it to 20 schools in the first edition. So, even from those twenty schools, we got about 3500 registrations. Through an assessment test, 300 students were selected to participate in the ML Summer School. We had got very good feedback from students who felt the content was really good and they learned a lot about machine learning, and they got exposure to many different areas or sub disciplines of machine learning, right from supervised learning to unsupervised learning, deep learning, reinforcement learning, probabilistic graphical models and so on. Building on this success, we decided to open the programme to schools across India.
This year, we want to look at maybe having a bigger class size. And again, we will have it open for all engineering students enrolled in any recognised institute in India. So, this edition will be very broad, and we hope to have a bigger class size so that more students can benefit from this programme. We've always been taking feedback from students to add more modules, to incorporate cutting-edge concepts within the summer school. This is a super dynamic field and in the last six months, generative AI and Large Language Models have become huge so we’re looking to incorporate some of that into the summer school modules so that students get an exposure to natural language processing, large language models, the latest architectures and the latest technology out there. So, we will incorporate more cutting-edge topics, such as large language models and generative AI into our content.
While making the workforce ready is one aspect, how has the customer experience evolved due to the surge in interest in machine learning technology?
Overall, AI and ML have been playing a major role in the industry to improve the customer experience. At Amazon, we use machine learning extensively to ensure that customers have the broadest selection of products to purchase from, at the best prices, and to deliver it as fast as possible, making it super convenient for them to buy products online. Amazon has been a pioneer in machine learning, and this is a journey we've been on for over 10 years now. We were pioneers in recommendation systems, building collaborative filtering techniques for recommending products. Based on the interests of customers that we've learned from past purchases and browsing history (on Amazon marketplace), we make recommendations to customers for products that they would most likely be interested in purchasing.
We have many experiences on Amazon, right from showing customers products that are regionally relevant. For example, customers searching for saree in the south may be interested in different types of sarees than somebody searching for a saree in the east or the west or the north of India. People have strong regional preferences and so, we show our customers products that may be more popular within their regions. We also learn about customers' proficiency based on the way they use our application and websites. For example, if a customer is issuing many search queries, they are likely to be more proficient, and we use that proficiency signal to rank widgets on the Amazon.in website. For customers with lower proficiency, we may show tutorials, onboarding tutorials, different language options, while for customers with higher proficiency, we may show them subscription widgets or our sponsored product ads and so on. We personalise the experiences for customers in many different ways. And of course, as I mentioned, we've been one of the early proponents of AI and machine learning to improve the customer experience.
There has been a lot of discussion about the ethical use of machine learning technology and making sure that the safety aspect is not ignored. What are the steps Amazon is taking in ensuring the ethical use of the technology?
As I mentioned earlier, we use machine learning primarily to improve the experience for our customers, to show them relevant content, to show them relevant products that they would be interested in purchasing, to improve the customer experience in terms of the broadest selection that they could find with millions of items at the best and most competitive prices and deliver to them at the fastest speed. So, our use of ML is really aimed at delighting customers and providing them with the best customer experience.