Subramanyam Kasibhat and the art of starting over

Subramanyam Kasibhat asked his team for confidential feedback and discovered that the boss had become part of the problem. Early in his entrepreneurial career, he had been learning quickly and expecting everyone around him to match his pace. They were struggling. His response had included anger and outbursts. The people working for him thought he needed to understand their circumstances and handle difficult conversations in private.

His first reaction was defensive. That detail gives the episode its bite. Asking for honesty is easy while the answers remain imaginary. Once colleagues supply the real thing, the invitation can suddenly look like a clerical error. Kasibhat eventually accepted that he could learn from criticism without first winning an argument about whether every part of it was fair.

He later put this uncomfortable memory into writing, tracing his entrepreneurial beginnings to 2000. Here was an engineer encountering a system that could answer back. He had requested information. Acting on it required changing himself. The account makes an unusually candid entrance to a career otherwise described through technology and company building.

What happens when someone who enjoys solving problems discovers that an old answer has stopped working? Kasibhat has repeatedly found reasons to reconsider his methods. Sometimes that has meant rebuilding software. Sometimes it has meant sharing authority. More recently, it has meant returning to the uncertainty of a new venture.

A partnership arranged on earth

Before the factory software, there was engineering. Kasibhat studied at SLN College of Engineering between 1992 and 1996 and worked as a design engineer at Texas Instruments India. Looking back after a reunion in 2025, he credited mentors, colleagues and friends there with helping shape his career. His account dwelt on the people and the chance to reconnect.

That technical history has a tangible footnote: he is named alongside Ashish Dharmpal Aggarwal as an inventor on an American patent application for an electrical system for a speaker and its control. Filed in 2009 and published in 2011, the application belongs to an earlier strand of work in audio electronics. Factory software was one destination in a broader interest in building products.

His enduring business partnership is with Savita Kasibhat, his wife and Vegam’s co-founder. Their marriage was arranged by their parents. In a 2025 reflection, he described a division of work: he concentrated on strategy, innovation and customer relationships; she led engineering, delivery and culture. These are substantial territories to negotiate across the same company and the same dinner table.

Savita’s reply supplied the domestic commentary that corporate biographies tend to misplace. Their parents, she joked, had managed to arrange a startup partnership as well as a marriage. She referred to the product roadmaps, debates and release cycles they now shared. The exchange makes the partnership feel like work done together, with room for disagreement.

One customer. Seven years. Then a rewrite.

A friend brought the couple a contract to develop software for a manufacturing plant belonging to a German company. They began without a manufacturing background. Seven years with that first customer preceded a rewrite of the software for a wider market. An assignment became an education, and the education eventually became a product business.

A seven-year commitment is hard to compress into the usual language of startup speed. It leaves plenty of time for a customer to expose the difference between an elegant demonstration and a useful working system. The eventual rewrite acknowledges another difficulty. Software shaped around one relationship needs a different architecture when other customers bring their own requirements.

By November 2023, Vegam was described as having 180 remote employees, mostly in India, and revenue approaching $10 million. Those figures describe the business at that time. Growth had come through selling to large customers with complex requirements, without outside funding. The patient early relationship had preceded an international software business.

Kasibhat argues that a method needs to make sense in its particular circumstances. That position is easy to admire in retrospect; it is harder to sit with while one customer occupies seven years. His career offers a concrete example of the time that independent judgment can require. The product eventually traveled. First, its builders had to learn.

For readers accustomed to the quick founding anecdote, this is a slower sort of beginning. The customer relationship carries the plot. The rewrite marks a decision to preserve what the founders had learned while changing the machinery through which other manufacturers would receive it.

A factory has to like the software

At ChemTECH in Mumbai in 2019, Kasibhat appeared on an AI and big-data panel with executives from Happiest Minds, NASSCOM, L&T Technology Services and Siemens. His explanation of a smarter factory focused on making decisions at the appropriate moment with useful information. It was a practical description of an ambition easily obscured by fashionable terminology.

Another discussion at the event addressed access: he argued that smaller companies could use available technology to compete with larger businesses without enormous investment. Technology had to earn a place in the work. The interesting question was what a business could do with it once the conference discussion was over.

Consider Vegam’s work with building-chemicals group MBCC. The company wanted to improve batch production, and the chosen starting point was the vOEE module. Kasibhat described years of collaboration with chemical manufacturers as preparation for the demands of that assignment. Batch production brought complexities the software needed to accommodate.

Adoption mattered alongside capability. The MBCC team wanted an application people could learn and then use independently after a short instructional period. A production manager spoke of workers liking the application enough to make it a friend. It is a pleasingly human specification for industrial software: the tool should be welcome when it turns up for a shift.

For a founder associated with customer relationships, this is where the job becomes specific. A sale leads into somebody else’s routines, expectations and existing equipment. The useful measure is what those people can do with the system once the visitors have gone home. A dashboard can be handsome and still be an unwelcome colleague.

The cost of being certain

Kasibhat returned to the subject of flawed judgment in April 2025. Working with younger teammates, he had noticed plans resting on untested assumptions and decisions made without enough questioning. He called the problem the Cost of Poor Analysis. He also explicitly included himself among the people who needed to improve.

His proposed habits were ordinary enough to be demanding: write down assumptions, challenge them, investigate causes, accept responsibility for outcomes and keep refining the analysis. There is little glamour in documenting an assumption. There is even less in discovering that the assumption belongs to the person who convened the meeting.

In a separate reflection on startup problem selection, he described how enthusiasm for a solution could obscure whether the underlying problem deserved solving. He asked other founders how they tested their assumptions. Product enthusiasm supplies energy; it also needs somebody willing to inspect the brief. His public writing leaves that question open for discussion.

The connection to his early feedback experience is an interpretation, but a persuasive one. Both accounts concern information that an enthusiastic decision-maker may prefer to disregard. In one case, colleagues had something difficult to say about his behavior. In the other, a cherished product idea needed examination. His advice asks the founder to remain available for correction.

“What assumptions am I taking for granted?”Subramanyam Kasibhat · April 2025

Back at the beginning, with company

By March 2025, Kasibhat was at NVIDIA’s GTC discussing AI tools for manufacturers. He described three purposes for attending: showing the company’s work, finding partners and learning from new developments. The mix of investors, business leaders, system integrators and technology executives offered a different audience from a factory implementation meeting.

In July 2026, he described a succession plan under way at Vegam, with the business moving toward its next leadership team. He and Savita were returning their attention to startup work through AI Intime. The ambition he expressed for Vegam’s next phase was to reach $100 million in scale. That was a goal, rather than an announced result.

AI Intime addresses enterprise adoption of generative AI. Its proposition emphasizes control over data, models, workloads and the timing of technical changes. The platform describes testing new technology against a business’s evaluations before adopting it. The resemblance to his manufacturing concerns is clear: a new capability needs to behave within an operating business.

Kasibhat has also described AI agents working alongside Vegam’s teams, with people directing their work across functions including development, support, marketing and sales. As the available tools change, somebody still has to define the task and judge the result. That presents another management question for an engineer who once had to reconsider how he led people.

Then there is the luggage. In June 2026, he was recruiting a junior mechanical or industrial design engineer to help develop a suitcase product in Bangalore. The assignment involved prototypes, printed parts, plastics, assembly trials and suppliers. For a founder discussing enterprise AI, it is a wonderfully physical list. A wheel still has to fit.

These pursuits leave him with several kinds of unfinished work: a leadership transition, an AI business and a product that must survive contact with manufacturing. The interest lies in watching an experienced builder accept another round of questions. Somewhere between the factory floor and the suitcase parts, starting over has become part of the job.