This edtech startup runs an online programme on Machine Learning and Artificial Intelligence for engineering grads and entry-level techies.
Srikanth Varma was a machine learning (ML) scientist at Amazon until last year. At the company’s Palo Alto office, and later in Bengaluru, he was tasked with the responsibility of building products for its advertising division, and also for hiring young talent in ML and artificial intelligence (AI) roles.
The hiring exercise made Srikanth aware of the huge gap between what academia taught and what the industry needed.
“I interviewed thousands of people while at Amazon and figured that institutes were teaching theory, and not applied sciences,” Srikanth tells YourStory. “There is no focus on real-life ML problems that the industry needs to fix,” he says.
In June 2017, Srikanth quit Amazon after five years, and moved to Hyderabad, where he taught an offline AI course free of cost to a batch of engineers and entry-level techies. Some didn’t even have prior knowledge in analytics, but ended up taking the course, and, later, went on to do their own case studies.
Srikanth says, “I tried to keep my experiment base wide enough. From fresh engineering grads to even Amazon/Google employees took the course. It went from the basics of programming to state-of-the-art deep learning.”
The success of his “teaching experiment” led Srikanth to believe that there was, in fact, a need for applied AI courses in India. The industry is investing heavily in deep tech and analytics, and AI alone is expected to create 2.3 million jobs globally by 2020, according to TeamLease Services. In India, the demand for “AI and ML specialists” is estimated to grow 60 percent in 2018.
But, there might be a demand-supply gap when it comes to skills.
“Most courses are taught by university researchers and they teach theory. There is a complete lack of mentorship. The course certificate is almost useless in the industry. We need concrete solutions.”
Hence, he set up Applied AI Course last September to offer engineering grads and tech professionals a one-year online programme in AI and ML. At at the end of it, students are guaranteed a job in AI/ML or data science roles in corporations.
Course with a ‘job guarantee’
Srikanth roped in Murali Krishna as his co-founder. Murali had previously worked at an edtech startup that assisted students with entrance exams. The duo started recording last September and launched Applied AI’s programme in March this year. So far, over 140 hours of course content (of which 15 hours are free) have been created. More are being added regularly.
The course videos address a range of questions such as ‘how to build an auto generated music system’, ‘how does Quora find similar questions’, ‘how does Netflix recommend movies’, and so on. Srikanth says, “We looked at real-world data sets and built solutions. We took a set of apparel data from Amazon and built a ML recommendation engine.”
Students are required to take 30 assignments during the one-year course. These are typically real-world industry issues they have to fix with AI/ML algorithms. They are even encouraged to put their source code up on GitHub (world’s largest community of coders and developers).
Srikanth reveals, “Some students are finishing their assignments in six months itself. One year is the outer limit we provide.”
The startup claims that it has already begun “placing students” in ML and data science roles in India and Japan as part of its job guarantee programme. Srikanth says,
“We want our success and our students’ success to be fully aligned. If we cannot place students within six months of their course completion, we refund the entire fee.”
Interestingly, the course fee was a matter of much deliberation. The co-founders looked at the average cost of engineering courses across the country and decided to price its programme at about a third of that figure. So, Applied AI is offering its course at Rs 25,000 + GST, which is what “an Indian middle class family sees value in”.
Srikanth says, “Our course completion rates have gone up every month since we launched, and 12 percent of the course-takers are from the US,” he adds.
Clients so far
While the startup refuses to reveal the current number of students, it claims enrolment rates are going up 10-15 percent every month. It counts a few engineering institutes and technology companies among its clients.
One such customer is global software corporation CA Technologies. Heena Raval, Senior Director at CA Technologies, tells YourStory,
“I have been very impressed with the customer centricity of the Applied AI team and how deeply engaged they were with us when we were working on the citizen data science programme for the ML/AI up-skilling of our engineering workforce.”
Educational institutes are benefiting too. BL Raju, Principal of Hyderabad’s ACE Engineering College, says, “We are glad to be associated with Applied AI for the past 10 months. It really helped our students to enhance their skills in AI/ML. About 108 students from our college are gaining both theoretical and practical knowledge.”
Challenges and path ahead
While on the surface it has been smooth sailing, Srikanth says course completion is the biggest challenge. At present, only seven to 10 percent of the students finish the entire programme.
“But that is an industry-wide issue. Even Coursera’s completion rate is less than eight percent. Students give up because either they don’t have time or they aren’t motivated,” he adds.
The 12-member startup is applying various measures to overcome this problem. For instance, it “starts calling up” individual students when their momentum drops. Sometimes, the handholding works. Also, the job guarantee programme serves as an incentive.
Applied AI Course continues to be bootstrapped and the founders are confident that it can run like that for two years. “I saved up enough at Amazon,” Srikanth reveals. “And our expenses here are not more than Rs 40 lakh a month. So, we are already making a profit.”
While it is not the only startup in this space, with the likes of Coursera, UpGrad, Springboard, etc. offering AI and ML courses, Srikanth reckons that those are “very academic” and do not focus on “applied stuff”. That, he believes, is Applied AI’s biggest strength.
Also, the moderate course fee could be a differentiator, with Applied AI's offering priced significantly lower than peer programmes.
The startup counts managers and engineers from Amazon, Microsoft, Samsung and others among its advisors even as it continues to be “inspired by” the Khan Academy model of simplifying complex subjects.
“What is our value proposition?” says Srikanth, “It is that we start with the fundamentals. No prior knowledge is required. We take advanced mathematics and make it as simple as possible.”