Deep-tech firms are still hiring because “hardware is driving demand”: Vayavya Labs HR

For months now, the tech industry conversation has sounded almost identical everywhere: layoffs, AI replacing jobs, slower hiring and shrinking teams.

But inside India’s deep-tech sector, the mood looks surprisingly different.

Companies working in semiconductors, automotive software and embedded systems are still hiring and, according to Girija Shetty, HR leader at Vayavya Labs, the reason is simple: the kind of work these firms do cannot be scaled or replaced as easily as mainstream software jobs.

“While we say software is eating the world, Hardware is driving the demand,” Shetty told People Matters.

That one line probably explains why a section of the tech industry still looks busy even while another part is cutting headcount.

Why deep-tech companies are not slowing down hiring

According to Shetty, many layoffs in the broader tech industry are happening in businesses that scaled rapidly for user growth.

“The companies laying off were scaling headcount for user growth,” she said. “Look at where the layoffs are happening mainly it is in consumer internet, fintech, edtech.”

Deep-tech companies work differently.

Instead of chasing app downloads or user numbers, they build systems tied to chips, automotive platforms and hardware products that often take years to develop.

That creates demand for highly specialised engineering roles such as:

• Linux device drivers
• RTOS development
• Embedded systems
• Virtual platforms
• Automotive software stacks

“We're not scaling user bases; we're solving engineering problems where each skill is unique and not easily fungible,” Shetty said.

And that demand is growing because hardware itself is expanding rapidly.

More connected devices. More chips. More AI-powered systems. More software-defined vehicles.

All of them need engineers who understand how software interacts with hardware in the real world.

India has engineers, but companies still cannot find the right ones

One of the more interesting parts of the interview was Shetty’s explanation of what companies actually mean when they say there is a “talent shortage”.

According to her, the problem is not the number of graduates entering the market.

“It's a precision problem pretending to be a volume problem,” she said.

India produces thousands of electronics and engineering graduates every year. But deep-tech firms often struggle to find candidates who can immediately work on complex production systems.

“A student who learned embedded C in a lab isn't ready for commercial automotive or semiconductor work,” Shetty explained.

The problem becomes obvious when engineers move from classroom exercises to real-world systems involving:

• Hardware constraints
• Timing deadlines
• Power budgets
• Field deployment issues
• Real-time debugging

“The issue isn't headcount but deployability,” she said.

That is why many deep-tech companies are now focusing less on flashy resumes and more on whether engineers can actually solve hard technical problems.

“Not just can they write code, but how do they debug? Can they think systematically? Can they explain tradeoffs?” Shetty said.

AI is changing hiring, but maybe not in the way people expected

At a time when generative AI tools are writing code, testing software and automating documentation, there is growing anxiety around whether software engineers themselves may become replaceable.

Shetty sees the situation differently.

“In fact, in our world, GenAI is raising the bar for what ‘good talent’ looks like, not lowering it,” she said.

According to her, AI tools work best when experienced engineers know how to review, audit and correct the output.

The challenge is not generating code anymore. The challenge is understanding whether the code will actually work inside complex hardware environments.

“Can you tell if the generated driver code will cause timing issues in a safety-critical system?” Shetty asked.

“Can you debug when the AI-suggested approach doesn't work in real hardware?”

In other words, AI may reduce repetitive coding work, but it is increasing the value of engineers who understand systems deeply.

Why ‘volume hiring’ may no longer work

Perhaps the sharpest observation from the interview was Shetty’s view that India’s tech industry is splitting into two very different worlds.

On one side is traditional “volume hiring”, the model that powered India’s IT boom for years by hiring large batches of generalist engineers.

On the other side is what she calls “precision engineering hiring”.

“GenAI and automation are killing that model,” Shetty said, referring to repetitive software work.

“If your value is executing a playbook, you're competing with tools.”

Deep-tech hiring works differently because companies are looking for people who can solve unusual, high-stakes engineering problems rather than execute repeatable tasks.

That is also why these firms hire more cautiously and invest heavily in training.

“You can't just add more engineers to an embedded software project and go faster,” Shetty said.

Unlike traditional IT services work, deep-tech projects often require long ramp-up times and years of accumulated expertise.

The next tech boom may look very different

The interview also hints at something larger happening across India’s technology workforce.

For years, mainstream software and application development dominated engineering aspirations. But the rise of AI, connected devices and semiconductor manufacturing is pushing attention back towards core engineering and hardware-linked roles.

According to Shetty, that shift could reshape career decisions for younger engineers entering the industry.

“Depth matters more than brand names or starting salary,” she said.

Her advice is unusually direct for a generation often focused on quick career growth and trending tech stacks.

“Work on hard problems close to the hardware. Learn how real systems work operating systems, drivers, real-time constraints.”

Because increasingly, the safest tech jobs may belong to engineers building the systems AI still cannot fully understand on its own.