TThe AI Generalist is a Myth

The most common job posting in Indian AI in 2026 reads, with slight variations, "AI engineer with experience deploying large language models in production." The same posting goes out from a hospital chain, a law firm, an agricultural inputs company, a school network, and a government department. The recruiters who write the posting believe, plausibly, that they are looking for the same person. The applicants who respond, mostly, believe they are qualified for all of these jobs. Both groups are mistaken in the same way. AI in a hospital, AI in a courtroom, AI in a farm, AI in a classroom, and AI in a government office are five different professions, not five verticals of one profession. The market has not absorbed this distinction yet, and the cost of the confusion is enormous.

¶Why the generalist illusion persists

The illusion that there is a single "AI engineer" persists because, at the model layer, the work looks similar. Pull a model. Fine-tune it on data. Wire it into an interface. Evaluate it. Ship it. The mechanics are roughly transferable across domains. The mechanics are also a small part of the job.

Most of the job, in any serious AI deployment, is not mechanical. It is the work of understanding the domain well enough to know what the model is being asked to do, what the cost of being wrong looks like, what the regulatory environment requires, what the users will actually accept, and what the failure modes will be in practice. That work is not transferable across domains. A person who is excellent at deploying AI in a hospital cannot, with a weekend of reading, become excellent at deploying AI in a courtroom. The substrate is different. The stakes are different. The judgement required is different.

¶A doctor with engineering literacy is not a medical AI engineer

The natural response is to say: fine, then the AI engineer should partner with a domain expert. The problem with this response is that the partnership rarely works. The engineer cannot fully internalize the domain in a part-time conversation, and the domain expert cannot fully internalize the engineering in the same way. The product gets built by the two of them shouting across a gap, and the artifacts produced reflect the gap.

What actually works is a third type of professional: the domain-AI hybrid, who is trained deeply enough in both the domain and the AI stack to make integrated judgements. The doctor who also knows how to evaluate a model, design a prompt, and interpret a hallucination rate. The lawyer who can audit a model's behaviour on case law and design a deployment that respects court procedure. The agricultural scientist who can fine-tune a model on local-language farmer queries. The teacher who can structure a curriculum around a tool whose limits they understand.

These hybrids are rare. There are perhaps a few hundred of them across all major Indian AI domains, against a need of tens of thousands. Producing them is the most important workforce decision the Indian AI ecosystem has not yet made.

¶The domain professional has the advantage

Here is the encouraging part of the diagnosis. The hybrid is much more easily produced by training a domain expert in AI than by training an AI engineer in a domain. The domain expert has spent years building the deep contextual judgement that the work requires. They have absorbed the legal, ethical, social, and professional context of their field at a depth that no AI engineer can short-cut. They need the engineering training, but the engineering training is, in a sense, the lighter half. A doctor with two years of serious AI training is a more useful medical AI professional than an AI engineer with two years of clinical exposure.

This inverts the usual pipeline of how AI workforce is recruited. The usual pipeline pulls engineers from generic programs and tries to layer domain context on top. The right pipeline pulls domain professionals, doctors, lawyers, teachers, civil servants, agriculturists, and trains them in the AI stack as a second skill. The domain professionals are also, conveniently, the people who care most about getting their field's AI right. They are the natural workforce for the work.

¶What the training pipeline should be

A serious "domain AI" training pipeline does not look like a generic data science bootcamp. It looks like a structured cohort of mid-career domain professionals, twenty doctors, twenty lawyers, twenty teachers, twenty civil servants, going through, over a year, the AI training that is calibrated for their domain. The mathematical depth is moderate. The engineering depth is sufficient to deploy. The domain depth is assumed and built upon. The cohort produces, at the end, professionals who can actually own the AI in their organizations.

This is structurally different from what most Indian AI training programs currently produce. Most programs produce junior engineers who can ship code. They do not produce senior domain-AI hybrids who can own deployment decisions. The latter is the bottleneck.

¶Bharath.CLUB as the natural home

This is one of the more specific reasons that a cross-domain professional community matters. The doctor learning AI cannot easily find their cohort inside a hospital. The lawyer learning AI is rarely in a firm that has more than one of them. The teacher learning AI is, often, alone in the school. The cross-domain community is where these professionals find each other, not so much for technical learning, which is solvable, but for the social texture of being in a cohort that takes the work seriously.

A doctor learning AI alone is a doctor learning AI slowly. A doctor in a cohort of ten other doctors and twenty cross-domain peers is a doctor learning AI on a curve that compounds. The community does not replace the technical training. It accelerates it, and more importantly, it gives the new domain-AI hybrid a place to belong professionally while they grow into the role.

¶The market will eventually figure this out

The market will, eventually, learn to write job postings that specify domain-AI hybrids. It will, eventually, pay them what they are worth, which is significantly more than either a generic AI engineer or a pure domain expert, because the combination is rare and creates outsized value. The pricing signal will reach the training pipelines, which will reorient to produce the hybrids. The whole correction will take a decade.

The Indian professional class can short-cut this by starting now. Mid-career professionals in any major Indian domain who add serious AI training to their existing expertise will, over the next five years, become among the most valuable workers in the country. The path is not glamorous. It is a year or two of effort on top of an existing career. The reward is being one of the few hundred hybrids in your field at exactly the moment the country needs tens of thousands of them. Bharath.CLUB exists, in part, to help that transition happen with peers in the room.