Last year, Infosys offered packages of up to ₹21 lakh for the role of specialist programmers. Then, HCLTech also joined in this month, offering salary packages of up to ₹22 lakh for AI-ready freshers.
This inflated the hopes of many skilled graduates seeking jobs. However, reality bites.
Most freshers entering Indian IT this year will still earn ₹3.5–4 lakh. A thin elite will earn 6x that. The gap indicates that companies are increasingly differentiating compensation by skill, with AI reshaping hiring patterns based on deliverables.
Neeti Sharma of HRtech firm TeamLease Digital sees the divide clearly.
“The candidates being paid higher are expected to bring skill sets that are relevant to the AI plans that the company has,” she tells AIM. “Skills such as designing AI workflows, data architecture and retrieval integrate AI into existing enterprise systems.”
Most freshers, she says, still walk into the same roles the industry has hired for decades. “Most freshers focus on support and basic development, using the available AI tools rather than building them.”
Sanchit Vir Gogia of Greyhound Research says AI tools have become universal to IT delivery.
“Developers are using GitHub Copilot. Test engineers are relying on GenAI to write and optimise test cases. Support teams are using AI to summarise logs, suggest fixes, and triage incidents,” he notes. “These tools are not being tested. They’re being used, actively and at scale.”
Gaurav Vasu of cognitive intelligence platform UnearthInsight agrees. “Teams are actively using GitHub Copilot for coding, refactoring, and test generation. ChatGPT and Azure OpenAI are being used for documentation, analysis, and agent assistance. Most large firms now run internal LLMs as secure enterprise copilots. Low-code tools with AI are being used to automate workflows across engineering and operations.”
So, What Has Changed?
What has changed is not coding; it’s evaluation.
“It’s no longer about how fast you can write clean code,” Gogia says. “It’s about how well you can review AI-generated outputs, tune prompts, build prompt workflows, and validate LLM behaviour.”
Vasu says the bigger shift is not the tools. It is what freshers are now being hired for. This is where salary premiums matter.
Sharma adds, “Freshers with practical AI skills such as prompt design, workflow automation, data prep, and building simple copilots are earning far more. Overall, AI skills are fetching approx. 35–40% salary premium.”
The highest premiums sit at the AI-native layer, where delivery changes structurally.
“Other high-demand skills include chunking strategies for RAG (Retrieval-Augmented Generation) systems, embedding optimisation, prompt templating, retrieval scoring, and access policy enforcement,” Gogia says. “There’s also a growing premium on LLMOps.”
LLMOps is heavily focused on governance, evaluating metrics, prompt drift, and QA of agent behaviour over technical deployment. “This is where risk lives,” he says.
Vasu sees the same hiring logic. “Freshers are not hired to build AI models from scratch,” he says. “They are hired for applied AI delivery skills that can be deployed immediately on projects.”
He highlights that roles involving prompt chaining, workflow automation, data preparation, output validation, basic agent building, and bias awareness will fetch higher pay. “Pure theory and generic certifications without delivery context do not attract premium pay,” he says.
These skills, he says, directly affect whether AI can be used safely inside real enterprise systems. “They help teams use AI productively inside engineering, operations, and business workflows.”
Engineers are no longer writing only deterministic code. They are designing systems that reason, retrieve, summarise, and trigger actions. “Other high-demand skills include chunking strategies for RAG systems, embedding optimisation, prompt templating, retrieval scoring, and access policy enforcement,” Gogia says. “There’s also a growing premium on LLMOps.”
“The pyramid is shrinking at the base,” he observes. “Repetitive junior roles are being absorbed by AI.”
“Setting up evaluation metrics, managing prompt drift, and running QA on agent behaviour in production” is now a full-time role, he adds. This skill shift is already reshaping the delivery pyramid. What remains are roles that supervise, orchestrate, validate, and govern.
“The fresher who knows how to work with AI tools, govern them, and explain their outputs is already leapfrogging those stuck in old delivery models,” he says.
Sharma sees similar trends when it comes to hiring. “Regular AI and ML fresher salaries range from ₹6–12 lakh,” she says. “Niche AI roles can pay ₹16–35 lakh, with specialised GenAI roles crossing ₹22 lakh.”
However, such roles are few and far between. “Most AI projects focus on chatbots, document automation, code modernisation, predictive maintenance, and internal copilots,” she says. “Very few involve training large models from scratch.”
Client spending reinforces this trend.
“The highest AI spending today is in IT ops, finance, customer service, sales, and manufacturing,” Sharma says. “Clients are paying for fewer incidents, faster invoice processing, fraud detection, higher conversions, reduced downtime, and productivity gains.”
No New AI?
Coforge CEO Sudhir Singh said during the company’s Q3 earnings call that when it comes to hiring AI-ready fresh graduates, unlike engineers from earlier generations, they have never really relied on certifications to prove their value.
“They come from an ecosystem in colleges globally, which is very hackathon-centric. They are used to being put into situations where they have to find a solution. Given how powerful AI is as a technology, that is the number one skill that you look for, and given the fact that the new cohort of engineers fresh from college has that in abundance, we have definitely not changed the talent catchment that we go after,” Singh explained.
This explains why most freshers still enter traditional streams. The industry does not yet need millions of AI engineers.
Also, hiring rhetoric almost always precedes reality, as actual hiring is based on client requirements.
“The highest AI spending today is concentrated in IT operations, finance, customer service, sales, and manufacturing,” Sharma says. “Clients are paying for measurable outcomes such as fewer incidents, faster invoice processing, fraud detection, higher conversions, reduced downtime, and productivity gains.”
These functions prioritise outcomes over model architecture. That pushes hiring demand for applied engineers, workflow designers, and governance specialists over pure researchers.
There is also a quiet shift in how engineers are evaluated.
“Engineers are now expected to work alongside AI systems, validating and governing outputs, rather than coding everything manually,” Sharma says.
Governance is now part of the job. “Most enterprises have now put clear governance in place,” Gogia says. “Shadow AI usage has been replaced with secure sandboxes, private LLM environments, and logging policies.”
Indian IT’s talent reset isquietly underway. It is not a hiring boom; it’s a sorting mechanism. Most freshers will still enter at ₹3.5 lakh and learn to work with AI tools, but only a small elite will design the systems.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.