Employees have to get on the AI train and go where it takes them: KFintech CTO

The workforce today has to get on the artificial intelligence (AI) train and go where it takes them, believes Nazish Hussain Mir, chief technology officer, KFintech, adding that the choice of the bogie still remains with the individual.

KFintech is India’s second largest registrar and transfer agent (RTA) for mutual funds and corporate issuers. An RTA is the back-office record-keeper that sits between a mutual fund (or a listed company) and its investors. When an investor invests in a mutual fund or any other asset class, they actually interact with the RTA and not the asset management company.

At the Global Fintech Fest (GFF) 2026, Mir told The Economic Times Digital that employees across KFintech’s verticals have adopted AI.


Asked if the organisation has a preferred large language model (LLM), Mir said KFintech keeps switching between models as newer versions emerge, depending on “whatever is the flavour of the month”.

Also Read: With AI handing down verdicts, India's lenders ask who signs off on it: Finarkein CEO

“If we talk about non-core functions such as finance, Claude today seems a better tool for us. The HR team has some ready-made custom tools from PeopleStrong. For engineering, it’s not uniform. Majorly, our teams use a combination of [OpenAI’s] Codex and [Anthropic’s] Claude. The UI-UX and product teams, too, use Claude Design.”

Mir also pointed out that KFintech has deployed an in-house AI Gateway to control how employees access and use AI models. The gateway acts as an orchestration layer through which all model and tool calls are routed, with a model registry defining which models employees are allowed to use.

“Everything passes through this gateway. We look out for certain keywords for compliance and regulatory aspects, then we understand which user [employee] is making a specific call and give them access to that specific model,” he said. KFintech built this control framework before deploying AI models at scale, Mir added.

He also said most of the company’s code is now AI-generated, although he did not have specific percentage figures.

The generalists and the specialists

While traditional engineering roles continue to make up the bulk of tech hiring, people with niche skills are increasingly in demand. Forward Deployed Engineers (FDEs), professionals who embed with a customer to build and adapt software for their specific needs, are one such category that is seeing a major hiring uptick, according to staffing firm Adecco.

At KFintech, the engineering teams are largely divided into two: generalists — employees who work horizontally across functions; and specialists — employees who work on specific verticals.

“There are generalist engineers in our organisation who use AI to build your application stack, the back-end stack… And we have specialist AI engineers, like FDEs. We call them to look at very specific problems. The generalist engineers can't get into such depth and are more horizontal. The specialist engineers go into the depth to solve challenges that no one has solved yet,” Mir explained.

Also Read: AI-led payments will need strong risk checks as trust remains critical: Amazon Pay India CEO

He also added that the number of generalist engineers is always more than the number of specialist engineers.

Even as AI increasingly generates the first line of code, fintechs in the banking, financial services and insurance (BFSI) sector have to remain mindful of regulatory requirements. Mir agreed, saying that because KFintech handles sensitive customer data, checks and balances need to be in place.

He added: “... We have a comprehensive set of test cases. Whatever code it [AI] generates, all the test cases have to pass. Otherwise, let’s say someone changes their mobile number in the portfolio; you cannot ask for another change in 10 days as mandated by SEBI. But what if, while automating this process, AI removed this 10-day cooling period? So, I need a set of entire test cases to identify it.”

On the business front, AI adoption has yielded results for KFintech, Mir said. The firm’s go-to-market (GTM) time — the time taken to bring a product or service to market — has improved by 30-40% as of September 2026, with greenfield projects seeing greater gains than brownfield ones, he added.

Pointing out the difference, he said, “Brownfield is a different scenario, because you have to look at legacy integrations and so on. But with [ChatGPT’s] Astra and [Claude’s] Fable 5.1, there have been tremendous improvements in brownfield projects also. They can one-shot your entire old code base and give it to you.”

Tier-II hiring still going strong

ET reported in April this year that the 2026 batch of engineering students from Tier-II and Tier-III cities was seeing a surge in placements, bucking an overall slowdown in entry-level IT/ITeS hiring. Private institutes that diversified beyond regular recruiters and focused on internships, pre-placement offers, and targeted upskilling in areas such as AI reaped the benefits.

For KFintech, too, the smaller cities continue to form a catchment area for attracting talent. “Our hiring is still continuing. We mainly operate from Mumbai and Bhubaneswar, but have offices in cities like Vijayawada and Bhubaneswar. We hire from such places because they have a good amount of talent that matches our different lines of business, and we will continue hiring because we want to have a trained workforce by the time clients onboard us,” Mir told the Economic Times Digital.

But what has changed for these freshly graduated hires is that they are no longer trained on code, said Mir.

He added: “The people that we hire today, you don't have to train them on code. While they might know the basics, what becomes most important is to teach them system design. You will have to teach them security, what is your harness, the guardrails, evaluation, etc.”

Mir is of the view that AI can generate code, but reviewing those pieces of the code is critical to understand the system design, layout, and architecture. “... How do these pieces integrate and go into your infrastructure? And did the AI architect it correctly or not, right? This is what you have to teach the people. So, what you teach will significantly differ. It's no more coding.”

Senthil Gunasekaran, the company’s chief business development officer (CBDO), explained how KFintech’s workforce has evolved alongside its growth in revenue. He said the company had about 6,000 employees and revenue of Rs 300 crore around five years ago. Its employee count has since increased to about 7,000, while revenue has grown to roughly Rs 1,300 crore.

“Today, we are roughly five times the revenue and close to about Rs 1,300 crores in terms of top line. So, five times growth has added only 1,000 employees. What has happened over a period of time is the number of resources in operations has remained more or less stagnant, but the headcount in technology has increased. Back then, we had 300 employees. Today, we have 1,400-1,500-odd employees,” he said, explaining the shift.