A forward deployed engineer (FDE), a term that comes from Palantir, which made the role its entire go-to-market strategy, is neither a consultant, nor a developer in the traditional sense.
The FDE sits with the client, understands what the business really needs, identifies what an AI tool can actually do, and then builds a bridge between the two.
It is a role that requires the ability to write production code, diagnose business processes and tolerate operating in environments where requirements are unclear and IT infrastructure is 20 years old.
The profile is rare. The market has priced it accordingly, with total compensation at mid-level running between $300,000 and $400,000 annually at the American AI firms that have popularized this role.
In India, TCS announced that it intends to build a team of up to 8,900 of them. This number deserves more scrutiny than it has received. CEO K. Krithivasan told Reuters the FDE corps would represent 1-1.5% of TCS’s workforce, putting the firm in direct competition with OpenAI, Anthropic and Microsoft for onsite AI implementation talent.
The ambition is coherent. The question is whether the supply of people who can truly do the job exists within the firm, or anywhere near it.
The FDE role emerged because AI has a specific and consistent failure mode. A model that works in a demo does not always work for the enterprise; this is typically because companies are complicated in ways that demos do not account for: fragmented data, legacy systems, processes designed for a different era, and users who have not been told why the new tool is better than the old one.
A 2025 MIT Media Lab report noted that 95% of Generative AI pilots fail, largely on account of flawed integration and a gap between available tools and the teams implementing them. The FDE is the person hired to close that gap.
In theory, this is exactly where Indian IT services firms should be able to compete. They have spent 30 years inside their clients’ environments. They understand legacy systems. They know where the data is buried and why integration is harder than it looks.
N. Chandrasekaran of the Tata Group put it plainly at a conference earlier this year: the real value is the context and understanding of every enterprise’s business and technology landscape. That is a fair description of what the FDE does, and it is also a fair description of what these firms have always claimed to do.
The difficulty is that understanding an enterprise and being able to rewire it around an AI system are not the same capability. The FDE role requires an engineer who can move between a business conversation and a production codebase in the same afternoon, who can tell a client their proposed use case is not a good use of AI, and who can build a working system in an environment designed by someone else.
The traditional IT services pyramid was not built to produce that combination of business judgement, technical depth and willingness to deliver an uncomfortable answer. Rather, it was aimed at delivering process adherence at scale, for which it produced IT engineers who were very good at executing a specification that someone else had written.
The FDE has to write the specification. Frequently, this professional must convince a client that the specification they arrived at is wrong.
This matters because AI without a proper business case is expensive decoration. According to Paraform, a specialist recruiter that tracks the category, FDE job postings grew 800% between January and September 2025, which tells you how many enterprises reached the end of a pilot run with no idea what to do next.
As demand for FDEs rises, their supply remains a constraint for IT firms. The role’s profile requires production engineering fluency, client-facing credibility, AI implementation experience and genuine domain understanding. This takes years to develop and cannot be manufactured quickly from an intake of fresh recruits, however large.
Which brings the latest quarter’s results into sharper focus. TCS’s headcount addition of 9,279 was its strongest in four years. HCLTech grew its ‘advanced AI’ revenue by 62% while cutting headcount by over 3,000.
These are firms navigating the same transition from different directions, and the FDE announcement is TCS’s most explicit statement of where it believes the transition leads: fewer people doing routine work, more people doing work that requires an intimate grasp of client businesses.
The honest question is how many of those new positions at IT consulting firms will be filled by engineers who can already do what the role demands and how many by engineers quickly retrained and reclassified so that they can be sent to client sites with a new job title. This distinction would matter to clients trying to deploy AI in a business that cannot afford to be someone’s training ground.
Consulting firms have historically billed by the hour. An engineer whose job is to make a model replace hours is an awkward thing to put on a rate card. TCS will report results again in October. While TCS may have 8,900 engineers called FDEs, what it would really need are 8,900 engineers whom clients would choose even if the title disappeared altogether.
This requirement has a way of separating the narrative from the reality.
The author is a technology consultant and venture capitalist.