What Engineers Actually Do When AI Handles the Code
Five Indian tech leaders explain how AI is shifting engineering roles toward judgment, architecture, and complex problem-solving rather than replacing engineers.
On Engineers' Day 2026, five senior technology leaders spoke to Deccan Chronicle about what the job looks like when AI handles routine execution. The consensus: AI has not shrunk engineering teams, it has moved them upstream. Engineers are now expected to frame better problems, judge when AI output can be trusted, design scalable systems, and keep learning as the stack shifts underneath them. Technical depth still matters. Judgment matters more.
What happened
Published on 15 September 2026, Deccan Chronicle gathered views from five senior figures across Indian technology companies on how engineering roles are changing as AI and automation mature. Their positions span cloud infrastructure, enterprise architecture, software development, and deep tech.
| Leader | Organisation |
|---|---|
| Ridhima Sawant | Orient Technologies Limited (Chief Transformation Officer) |
| Joseph Sudheer Reddy Thumma | Magellanic Cloud Limited (Chairman and MD) |
| Tarun Dua | E2E Networks (Founder and MD) |
| Arun Kumar Sasidharan | SVP, Enterprise Architecture |
| Murali Swaminathan | Freshworks (CTO) |
What the leaders actually said
Tarun Dua of E2E Networks put it most directly: “AI hasn’t reduced the need for engineers, it’s raised the bar for what a good one does. The job is moving from writing every line of code to deciding what’s worth building, judging when AI’s output can be trusted, and finding the ideas AI alone won’t get you to.” He also noted that building India’s compute layer from scratch is generating more engineering work overall, just work that demands constant relearning.
Ridhima Sawant framed the change as a shift in the quality of questions rather than the volume of output. “AI can accelerate execution, but judgement, curiosity and a willingness to challenge established approaches will determine how effectively it creates value.” Her point is that the hard part is now deciding what should be built and why, not how fast it ships.
Joseph Sudheer Reddy Thumma, whose company works in drones, e-surveillance, and AI-led systems, described a widening responsibility set. Engineers in operational environments must make technology reliable, scalable, and relevant to real-world needs, not just technically advanced. He called the combination of technical depth, judgment, and continuous learning the defining trait of the next generation of engineers.
Arun Kumar Sasidharan put it in value-chain terms: as AI absorbs repetitive tasks, engineers move toward complex problem-solving, architecture, innovation, and strategic decisions. He was also clear about the ceiling: “Technology cannot replace human judgement, the ability to understand context, weigh competing priorities, and recognise the implications of a decision beyond the data.”
Murali Swaminathan of Freshworks echoed the theme: future-ready engineers need “a mindset of constant learning, experimentation, and adaptability,” not just stronger technical credentials.
Why it matters for teams shipping AI products
If you are hiring, managing, or working alongside engineers right now, the practical implication is that seniority signals are shifting. A developer who can prompt an AI coding tool and ship fast is table stakes. The scarcer skill is someone who can evaluate whether the output is trustworthy, catch the edge cases the model missed, and make the architecture call that keeps the system maintainable two years from now.
For businesses integrating AI into their products or workflows, this has a direct consequence: the cost of bad judgment has risen. When AI accelerates output, mistakes also scale faster. That is why AI integration work that looks straightforward on the surface often turns out to require significant human oversight at the design and evaluation stages, not just at deployment.
The infrastructure angle Dua raises is worth watching too. India’s compute buildout is creating demand for engineers who understand the hardware and networking layers, not just the application layer. That gap is not filled by AI tooling; it requires people.
Our take
These quotes read as genuine rather than defensive. Nobody here is claiming AI is overhyped or that engineering jobs are safe regardless. The honest version of the argument is: AI is a force multiplier, and multiplying bad judgment at speed is worse than moving slowly with good judgment.
From where we sit building and automating things for clients, the bottleneck we see most often is not writing code. It is the upstream work: defining what the system needs to do, choosing the right approach, and knowing when a vendor’s AI feature is actually solving the problem versus adding complexity. That work is harder to automate than people assume, and it is what separates projects that deliver from ones that drift.
The continuous learning point is real but often underestimated in practice. The stack is not just evolving; it is changing fast enough that last year’s best practice can be the current antipattern. If your engineering team is not allocating structured time for that, you will notice the gap, just later than you would like. You can see how we approach this balance across our client projects where AI and human oversight work in parallel.
What to do about it
- Audit your team’s skill mix: identify who can evaluate AI output critically, not just generate it.
- Build in review stages where a human checks architecture and edge-case decisions before AI-assisted code reaches production.
- Allocate explicit time for relearning. Monthly, not annually.
- When assessing new AI tools, ask what judgment calls the tool still requires of the user, and whether your team has those skills.
The engineers who stay valuable are the ones who treat AI as a collaborator to interrogate, not a black box to trust.
Frequently asked questions
Is AI replacing software engineers?
According to senior tech leaders speaking on Engineers' Day 2026, AI is not reducing demand for engineers but is changing the job. Engineers are moving from writing code to making architecture decisions, evaluating AI output, and solving complex problems that AI alone cannot reach.
What skills do engineers need in an AI-driven workplace?
The executives quoted highlight judgment, continuous learning, technical depth, and the ability to question assumptions and understand context. Adaptability and a willingness to keep relearning as the tech stack changes were cited as non-negotiable.
How does AI change the engineering role in practice?
AI handles repetitive execution tasks, freeing engineers to focus on problem definition, system architecture, technology evaluation, and strategic decision-making. The expectation is that engineers move higher up the value chain, not that they do less work.
Is India's tech sector creating or cutting engineering jobs because of AI?
Tarun Dua of E2E Networks said India is building its compute infrastructure layer from scratch, which is creating more engineering work overall, but requiring new skills and constant relearning rather than the same roles as before.
