As enterprises move from generative AI experimentation towards agentic applications, the skills required to build and deploy these systems are changing faster than traditional computer science curricula can keep pace. The challenge, however, is no longer simply about teaching students how AI models work. It is increasingly about giving them the context to apply AI across data, business processes, workflows, governance and enterprise environments.
That is the gap that UPES and Salesforce are attempting to address through a co-created AI-integrated B.Tech programme, which combines academic learning with hands-on exposure to enterprise technologies and AI platforms.
In an exclusive interaction with Express Computer, Prof. (Dr.) Abhishek Sinha, UPES, Sridhar Hariharasubramanian, Senior Director, Salesforce India, and Sangeeta Giri Gundala, SVP & Chief Operating Officer – South Asia, Salesforce, discuss why enterprise AI is changing the definition of job-ready talent, how the curriculum is being designed around experiential learning, and why governance and responsible AI need to become part of technical education.
The problem is a context gap, not a content gap
According to Sinha, traditional computer science education does not necessarily suffer from a lack of AI content. The bigger issue is that students often do not get the opportunity to apply what they learn in an enterprise context.
“If you look at the curriculum of any BCS programme, not just data sciences, any specialisation, whether in India or in any jurisdiction, the curriculum is very similar. So it’s not the question of content gap. The gap which we are trying to bridge is the context gap.”
The approach therefore focuses on experiential and participatory learning rather than simply adding another AI course or specialisation. “We are not trying to teach them what is agentic AI in the classroom. What we are trying to give them is a platform where they can actually utilise these technologies and learn by doing it.”
Sinha argues that access to information itself is no longer a differentiator for universities. Students can access technical knowledge through AI platforms and other digital resources. The role of education, therefore, is increasingly to provide an environment where that knowledge can be applied.
The collaboration follows what Sinha describes as a “Triple C” model — co-create, co-deliver and take students from classroom to corporate. The programme includes a co-curated data sciences specialisation, hands-on exposure to enterprise technologies, faculty training and a centre of excellence focused on agentic platforms.
Enterprise AI needs more than coding skills
Sridhar says the skills gap goes beyond traditional programming and model-building capabilities. “A lot of the courses are still very technically oriented. They focus on things like coding skills or building models and so forth. While that is definitely an integral part of it, that’s not the only thing.”
Building an enterprise AI application requires an understanding of data, integrations, business processes and workflows, alongside data security and privacy. “These are ultimately what is required to put up a real-world, enterprise-ready application as we call it. And I think those are the components which a lot of the courses today that you find in academia are generally missing,” he adds.
This is also influencing what the programme teaches. While agentic AI forms a major component, Salesforce’s focus extends beyond the technology itself to the foundations required to make AI useful in an enterprise.
Sridhar says the curriculum needs to cover “data, business processes, workflows” and governance, arguing that context can be as important as the underlying model’s capabilities. “The model capabilities are fairly advanced today. We don’t need even more advancements because with every advancement in the model, it is not like the enterprise applications are getting any better.”
From learning about agents to deploying them
The programme is designed around the principle of “hands on the keyboard from day one”, according to Sinha.
While students continue to receive the theoretical foundation of a B.Tech programme, the specialisation brings in hands-on technology exposure, deployment, certifications and participatory learning. Faculty members also undergo training as part of the programme. “The training part of it, the curriculum part of it, the deployment part of it, and then the tools which are available — which is not just limited to one or two tools — and then the certifications which are provided by Salesforce – these are the four-five buckets.”
The objective is to ensure that students do not simply understand AI agents conceptually but have experience deploying them.
Salesforce’s Sridhar adds that direct exposure to enterprise platforms is intended to make graduates more familiar with the environments in which they are likely to work. “Think of this as an engineering student learning about an agent in the classroom and never getting an opportunity to deploy an agent. This kind of collaboration gives that platform where they can actually go ahead and deploy an agent.”
Responsible AI becomes part of the curriculum
As AI moves into enterprise decision-making, both executives stress that technical capability cannot be separated from governance.
Sridhar says trust has always been central to Salesforce’s approach, but responsible AI becomes more complex because models can be probabilistic and can produce biased, toxic or incorrect outputs. “Models are known to be confidently wrong. The answer that it gives you looks really great, but that doesn’t mean that it is right all the time.”
Enterprise applications, however, often require deterministic and explainable outcomes. He gives the example of a financial services application making a credit decision, where a customer needs to understand why an application is rejected.
“The governance aspect in terms of the audit trail, how the decision was made, what kind of reasoning logic was used — you as a consumer have the right to go and ask the bank… you have the right to know why.”
For Sinha, governance also needs to extend beyond technology into education itself. UPES has established an AI transformation office covering areas including responsible AI usage, legal and ethical considerations and governance. He says the university has trained all its faculty members on AI basics and is extending AI literacy across its student population.
Building India’s AI workforce
For Sangeeta Giri Gundala, the collaboration comes at a point when India needs to expand its AI talent ecosystem alongside enterprise adoption.
“We believe India is at a defining moment in its AI journey. The conversation is no longer about whether AI will transform industries, that’s already happening. The bigger question is whether we can build the talent ecosystem that enables India to lead that transformation globally.”
She argues that the emergence of agentic enterprises is changing the skills expected from graduates. “As organisations evolve into agentic enterprises, the demand for talent that can work alongside AI, harness trusted data, and solve complex business problems will only grow.”
That requires a different relationship between academia and industry, she says. “This collaboration represents a new model for higher education, where industry doesn’t just advise academia, but actively co-creates curriculum, co-delivers learning, and co-builds the future workforce.”
She sees the model as potentially scalable across higher education, with deeper partnerships between industry, academia and government needed to prepare students for AI-driven roles.
For Sinha, however, success cannot be defined only by placement numbers several years from now. Given the speed at which AI is evolving, the immediate objective is to ensure students receive access to the best available learning and technology experience. “The time a university provides the best kind of learning possible in a specific area, automatically the by-product is that the talent pool is created.”
Sridhar similarly sees the opportunity in positioning India as a source of global AI talent, but says employability and responsible technology use will remain important measures of whether such programmes create lasting value.
The underlying message from the three executives is that preparing for the agentic enterprise requires more than adding AI to an existing syllabus. The next generation of technology professionals needs to understand how AI connects with data, business processes, enterprise systems, governance and human decision-making. As the technology continues to change, that ability to apply knowledge in context may become more valuable than knowledge of any individual AI tool or model.