Every few years, an obscure job title from one company becomes the industry's shorthand for what it's actually worried about. This year, that title is Forward Deployed Engineer.
Palantir popularised the term over a decade ago for engineers embedded inside client organisations, building software against a client's own data and constraints instead of shipping something generic off the shelf. Now, the rest of the industry is catching up. OpenAI has launched a dedicated deployment company backed by more than $4 billion, with around 150 experienced forward deployed engineers (FDEs) and deployment specialists coming in through its acquisition of Tomoro. The demand is visible in the numbers too: FDE job postings have risen 729% year-on-year, from 643 in April 2025 to 5,330 in April 2026.
India is seeing the demand too. Enterprise AI startup Wonderful has launched India operations and plans to hire up to 1,000 Forward-Deployed Engineers. LTM, meanwhile, has announced AI1000, a programme to develop 1,000+ AI-certified engineers, including FDEs.
A FDE is not a model researcher, and the job is different from a standard software or machine learning role too. FDEs don't train models. They take models that already work and make them work inside a specific company's environment: its data, compliance requirements, legacy systems and workflows.
Take a bank that wants an LLM to help loan officers process applications faster. A standard engineer can build an integration against clean sample data. An FDE has to understand the bank's legacy systems, privacy rules and the informal decisions loan officers make, then build the technology, guardrails and workflow around them. The job is to make AI work inside the mess, not around it.
McKinsey's 2025 State of AI research found that only about 6% of organisations qualify as AI high performers, and these organisations are 3.6 times more likely than others to pursue transformative change with AI. The opportunity is therefore not simply to build better models, but to redesign how businesses actually work with them.
{{/usCountry}}McKinsey's 2025 State of AI research found that only about 6% of organisations qualify as AI high performers, and these organisations are 3.6 times more likely than others to pursue transformative change with AI. The opportunity is therefore not simply to build better models, but to redesign how businesses actually work with them.
{{/usCountry}}That makes the FDE role unusually demanding. It requires AI and LLM fundamentals, backend engineering, cloud and system design, enterprise integration and security. But technical depth alone is insufficient. FDEs also need stakeholder management and customer communication and, above all, the ability to sit with a stakeholder who can't articulate their own problem and turn that ambiguity into something that ships.
This is where India's position becomes particularly interesting. Indian engineering built its global reputation on integrating complex enterprise systems and solving specific problems for specific customers at scale. But the IT services era largely rewarded efficient execution against defined requirements. The forward-deployed model builds on that foundation and also demands more: business judgment, technical depth, and comfort working through incomplete problems in real-world systems.
Nandan Nilekani has described the gap between AI capability and enterprise adoption as a “deployment gap”, saying that “technology is far ahead of its deployment.” He has also argued that India can become the “AI use-case capital of the world”, built around mass-scale implementation rather than simply building foundational models.
The opportunity is significant, but there is a skills problem. A Scaler-CMR study of 400 experienced software engineers and tech recruiters found that 89% of engineers believe they are AI-ready, but only 19% are deeply engaged in building AI/ML systems; 86% of recruiters say they struggle to find AI-ready talent. Our conversations with enterprises hiring for AI deployment roles have reinforced this gap, with demand strong enough to prompt the evolution from an FDE specialisation to a dedicated, full-time programme.
But this is more than a talent gap, it reflects a shift in engineering itself. The field is beginning to split into two distinct tracks: One focused on building AI, and another focused on deploying it. For the last few years, the conversation around AI has largely been about building better models. The next phase will be about making those models work inside real businesses.
The first track will continue pushing the frontier of models, infrastructure and core technology. The second will take that technology into real organisations: understanding the problem, integrating systems, managing risk, working with stakeholders and ultimately delivering an outcome. As organisations move from experimentation to production, this second track will become just as critical.
For engineers, this creates a new path to build careers at the intersection of technology and business. The Forward Deployed Engineer is one of the clearest expressions of that path, combining technical depth with customer context, problem-solving and ownership of deployment.
The starting point need not be a narrowly defined “AI engineer”. Strong software engineers are an obvious fit, but backend and full-stack engineers, data engineers, cloud and DevOps professionals, solutions architects and other technically adjacent professionals can also make the transition. The learning curve may be longer for those coming from adjacent roles, but strong engineering fundamentals provide the foundation on which AI fluency, enterprise integration and business problem-solving can be built.
The real test of the next phase of AI will not be how powerful the technology becomes, but how effectively businesses can put it to work. That is where the next generation of engineering talent will matter most.
(The views expressed are personal)
This article is authored by Amar Srivastava, CEO-online & group CPO, Scaler.