India has less than 2,000 senior engineers who can build core AI products

India has less than 2,000 senior engineers who can build core AI products

Technology product and IT services companies reported training thousands of employees in artificial intelligence (AI) skills over the past year, yet India is in the midst of a talent shortage in core AI products and services.

These technology companies are also investing heavily in upskilling their employees. Cases in point: Tata Consultancy Services recently said it had trained 350,000 employees in AI skills; Infosys said eight out of 10 employees are AI-ready; Accenture is upskilling 250,000 employees in AI; HCLTech has trained 50,000 employees; Microsoft India plans to skill 2 million people in AI by 2025, and IBM has globally committed to train 2 million people in AI by 2026.

Despite these lofty numbers, the active pool of senior AI engineers who are actually building core AI products and services in India is less than 2,000, according to data from specialist staffing firm Xpheno, accessed by Moneycontrol. Core AI products and services include AI-based apps, tools and platforms, to name a few.

The total active accessible Indian talent with the minimum required skills, experience and exposure to core AI is a little under 21,000, Xpheno said, referring to the segment of upcoming AI engineers.

This is quite a contrast to the 650,000-700,000 people that tech industry body Nasscom claimed are trained on AI across top-tier tech companies.

Then what explains this dichotomy? What are these thousands of people exactly getting skilled in?

Experts told Moneycontrol there are other aspects of AI skilling – beyond building core technologies – that are more geared to embedding and integrating AI to improve daily tasks and productivity. This training is not restricted to engineering talent but is also about skilling other functions such as marketing and even HR.

Levels of AI skilling

“The (AI) training that people are doing is more mass market,” said Jagdish Mitra, a tech veteran and board member of the non-profit National Skill Development Corporation. “And that’s not what will be valuable. What will be valuable is process-specific, application-specific and industry-specific skilling that will make a difference to how the skills are being used. It has to be more specific, otherwise it’s like a man walking around with a hammer trying to find a nail.”

Mitra added that even then, the current way of upskilling too will be important in the long run as it will create a fundamental layer of basic AI knowledge among people.

According to Sangeeta Gupta, senior VP at Nasscom, apart from core AI engineering, there are two key segments of AI training in companies.

Firstly, the skilling undertaken by companies such as TCS and Accenture, which involves upskilling employees who may not build the next AI tool or platform but will be involved in embedding AI in existing solutions like app development or software testing applications.

The other category of AI skilling is for people in companies to use for inhouse or work-related functions.

“So that will be, say, the marketing teams within the technology companies or BPO companies, and how they will teach each of their employees to leverage these tools to provide better outcomes and efficiencies,” Gupta told Moneycontrol.

What are these AI skills exactly?

Skilling engineering talent, especially in IT services companies, Mitra explained, will involve learning specific technologies pertaining to AI.

“At the heart of it is obviously data science, but there are other things like machine learning, natural language processing, opticals and so on… these are specific skills, which not all engineers are trained in. They (engineers) do different things. For instance, a testing guy or development guy or a Java development guy may know Java but will not know how to apply machine learning to it all out. So those are things we are teaching people,” Mitra said.

This boils down to engineers also becoming future-ready to build their careers amid AI disruption. A developer or a testing engineer will now also be skilled in machine learning, computer vision and more.

Gupta added that within India, machine learning (ML), Python, and SQL skills dominate the current environment, apart from generative AI skills. For generative AI, engineers are being upskilled for better understanding of foundational model tools such as OpenAI’s ChatGPT, Google’s Gemini, Meta’s Llama 2, and Microsoft’s Copilot.

AI demand and opportunity for India

Citing a recent Nasscom and BCG report, Gupta said demand for AI talent will grow 15 percent year on year; and there is already a demand-supply gap of about 51 percent for niche skills that will help build core AI.

According to Prasadh MS, head of workforce research at Xpheno, as a predominantly services market, India is all set and waiting on the sidelines for the spike in mainstream consumption of GenAI and AI in general.

“Core engineering of and for AI remains concentrated with high niche micro-talent pools operating outside India. The core engineering of AI has minor footprints in India. Global AI products looking for service and support skills in India are the first major wave the tech sector will ride in the AI domain,” he told Moneycontrol.

Krishna Vij, business head-IT staffing at TeamLease Digital, said that even though deals and demand are showing for IT companies, they continued to heavily invest in AI training because they’re not able to get the required talent from the market.

“Also, even if companies are looking at reducing their budget towards employee cost, they’ve kept a specific budget and are okay to shell out 25-30 percent more to identify the right talent for AI kind of role,” Vij told Moneycontrol.

Global Captive Centres (GCCs) too are at the forefront of driving AI talent demand, especially as many of them predominantly develop products in-house.

“In the last two years we’ve seen that GCCs have been the aggregator in terms of hiring because IT services companies were not hiring to the tune of what was expected to. So yes, GCC of course pays slightly more premium because you’re directly working with the projects,” Vij added.