AI to Replace Engineers by 2025—But Not in India

It’s 2025—are software engineers on their way out? Not just yet. While global tech leaders predict AI will replace coding jobs, Indian founders believe otherwise.

Recently, Salesforce chief Marc Benioff announced that the company will pause the hiring of software engineers in 2025, citing a 30% productivity boost from AI tools like Agentforce. Meta CEO Mark Zuckerberg believes AI will take on mid-level engineering tasks by 2025.

Meanwhile, Abhishek Upperwal, founder of Soket AI, doesn’t buy the 2025 theory. “I don’t think AI will completely replace mid-level engineers, at least not by 2025. Even OpenAI is still hiring web developers with salaries exceeding $300k a year,” he told AIM.

He added that Soket has been using Cursor for a while now, and it has definitely boosted productivity. “That said, we’ve learned that blindly relying on these tools can waste a lot of time as they’re prone to errors,” he said.

Upperwal, however, agreed that AI is likely to reduce the demand for mid-level software engineers, as tools like Copilot and Cursor can handle many repetitive coding tasks efficiently. “They’re great for things like generating code snippets, automating workflows, and even suggesting fixes,” he said.

Unperturbed by the global chatter, Puneet Jindal, CEO of Labellerr, said,  “We are still a decade away from replacing mid-level engineers,” arguing that AI leaders fuel this debate not just out of optimism but also to keep their company valuations unrealistically high.

“You will need mid-level engineers; it’s just that their productivity will be amplified with time. Those engineers who miss upgrading themselves will fall behind,” he added. 

Tony Mauro, chief technology officer of Plasmatic Technologies, made it clear that mid-level engineers are indispensable in the tech ecosystem. “They fix bugs, troubleshoot issues, and handle the less glamorous aspects of product development. They’re also the pipeline to future senior engineers, the ones who’ll step up when AI tools aren’t working as expected or when an outage brings operations to a halt. Without them, who’s left to debug the AI?” he said. 

Mankaran Singh, Founder of FlowDrive, commented on the complexity of replacing humans with AI. “It would take a long time to fully remove any human from the loop. Replacing humans requires very strong reasoning capabilities, which LLMs are not very good at as of now and will require some more breakthroughs in the field,” he said.

“AI will not replace mid-level engineers entirely but will augment their capabilities. In the worst-case scenario, it may reduce the head count. For instance, if we require 10 developers currently, in the future, we may need only 5-7,” said Santhosh HS, AI engineer at TCS. 

Global Leaders Think Otherwise

“We believe, by 2025, we may see the first AI agents join the workforce and materially change the output of companies,” Altman wrote in a recent blog post. Interestingly, though, Altman is also seeking to add 158 more employees to his team, according to the company’s careers portal. OpenAI is specifically seeking over 90 new employees for its research and engineering teams.

In a recent podcast, Benioff said that AI tools like Agentforce have significantly increased the company’s productivity. “We have increased productivity… to the point where our engineering velocity is incredible. We will have fewer support engineers next year because we have an agentic layer,” he added. 

Despite the reduction in support staff, Salesforce is planning to expand its sales team by 1,000 to 2,000 employees to communicate the value of its AI solutions to customers. 

Similarly, Meta CEO Mark Zuckerberg, in a recent podcast with Joe Rogan, said, “Probably in 2025, we at Meta, as well as other companies that are working on this, are going to have an AI that can effectively be a sort of mid-level engineer at your company that can write code,” Zuckerberg said.

He added that initially, running AI for these tasks will be expensive, but efficiency will improve over time. “And once you have that, in the beginning, it’ll be really expensive to run, but you can get it to be more efficient,” he explained. Meanwhile, according to recent reports, Meta plans to reduce its workforce by approximately 5%, targeting the company’s lowest-performing employees.

Klarna CEO Sebastian Siemiatkowski recently revealed that the company, known for its “buy now, pay later” services, has not hired any new staff in the past year, resulting in a natural reduction of its workforce. Klarna’s employee count has decreased from 4,500 to 3,500. 

In a recent post on X, Siemiatkowski clarified that AI was a big reason behind not hiring. “As we leaned in and explored the capabilities of AI, we realised a good portion of the kind of work that was done at Klarna will become redundant over time, some things faster, others slower,” he said.

Upskill Yourself

While AI is transforming the landscape by automating repetitive tasks, optimising workflows, and shifting the focus of mid-level engineers toward higher-value responsibilities, some believe, in the near future, the demand for engineers will be driven by those who can effectively collaborate with AI tools, using them to enhance productivity rather than seeing them as competitors.

“By leveraging AI, our engineers can focus on high-value tasks, while AI handles routine work like code suggestions and debugging,” said Ankush Sabharwal, founder of Corover.ai. He further added that this approach has not only reduced development time but also enabled their teams to deliver more robust solutions faster.

Siemiatkowski said that ever since ChatGPT came around, they have encouraged everyone at Klarna to learn how to apply and use AI. “There is massive adoption, and I believe this to be our most successful upskill program ever,” he said.

“To stay relevant, engineers will need to focus on high-complexity skill sets. These include optimising algorithms for CUDA [or any other accelerator], refining synthetic data generation, and mastering new languages and frameworks—things AI isn’t immediately equipped to handle,” said Upperwal.

Atishay Jain, founder of Hyperke.com, reiterated that prompt engineering and the ability to leverage LLMs to write quality code quickly are among the skills engineers should have. “They will also need to familiarize themselves with system design, architecture principles, and problem-solving skills,” he added.

Similarly, Santhosh said that AI tools are already being integrated into workflows to enhance productivity. These tools have streamlined several processes, from code generation and debugging to system optimisation. For instance, platforms like GitHub Copilot, Microsoft Copilot, and Azure AI services assist engineers in writing cleaner code more quickly.

 “This has allowed teams to focus more on critical thinking and the creative aspects of their projects, reducing time spent on mundane tasks,” he added.