AI, Automation, and the Future of Software: A Conversation with Madan Thangavelu
Madan Thangavelu has spent years in the trenches of high-scale engineering, building Uber’s core infrastructure and leading a global team of engineers. But if you ask him what the future of software looks like, he won’t just talk about frameworks or languages. He’ll talk about AI, automation, and a fundamental shift in what it means to be an engineer, a builder, and a decision-maker in an AI-powered world.
Indie Hackers sat down with Madan to dig into his outlook on AI’s impact on software development, leadership, and the future of work.
AI: The New Compiler, But More Than That
“AI tools today are like the compilers of the past,” Madan says. “Back then, writing machine-level code was an art. Then compilers came along and let us focus on higher-level problems. AI is doing the same thing for software engineering now. But the difference is that AI doesn’t just translate from one language to another—it can generate entire solutions, rewrite how we build software, and push engineering into a more strategic domain.”
He points to tools like Replit and Cursor, which are reducing friction for developers and making coding more accessible. But the real change isn’t just in writing code—it’s in how software gets built and who gets to build it.
“In the next five years, we’re going to see a shift where AI-assisted development becomes the default. Writing boilerplate code manually will seem as archaic as writing assembly language today. This changes the role of engineers from pure coding to higher-level problem solving. It also changes the profile of who can contribute. The next wave of successful engineers won’t just be the best coders—they’ll be the best at leveraging AI to achieve outcomes.”
The Shift: From Coders to Engineers Who Lead
If AI can generate code, does that mean fewer engineers? Not exactly. It means the role of an engineer evolves.
“The engineers who thrive in the AI era will be the ones who understand the business, the product, and how to align technology with company goals,” Madan explains. “In the past, only senior engineers needed to have that level of insight. Now, everyone does. AI is removing the barrier to entry for writing code, but it’s also raising the bar for decision-making. If you’re just executing on well-defined tasks, AI will outperform you. If you’re defining the strategy, making trade-offs, and thinking holistically, you’ll be irreplaceable.”
The differentiator? Soft skills and high-level decision-making.
“Writing code is becoming commoditized. But debugging? Understanding trade-offs? Leading teams? These are still rare and valuable skills. The best engineers will be the ones who can translate business needs into technical execution. They’ll be the ones asking, ‘What’s the right thing to build? How do we align technology with the company’s future?’ That’s where AI can’t replace human intuition—at least not yet.”
AI in DevOps: Smarter Monitoring and Testing
DevOps and CI/CD workflows are ripe for AI-driven improvements. One of the biggest pain points? System monitoring at scale.
“High-cardinality monitoring—tracking metrics across different dimensions like users, regions, or platforms—is really tough,” Madan explains. “AI can analyze anomalies more accurately than traditional monitoring tools. Instead of engineers sifting through logs and dashboards, AI can detect patterns and point directly to the root cause. This fundamentally changes how we respond to system failures and performance issues.”
Another big leap is in automated testing. “We’re moving beyond hardcoded end-to-end tests. AI tools like OpenAI’s Operator can simulate user actions in a browser, making testing more dynamic and adaptive. Instead of manually updating test cases, AI will generate and refine them in real time. This means software will get shipped faster, with higher quality, and with fewer engineers dedicated to QA.”
AI Ethics: The Biggest Blind Spot
As AI becomes embedded in decision-making, ethical concerns are becoming harder to ignore.
“The problem with AI isn’t just bias—it’s explainability,” Madan says. “If an AI system rejects your loan or makes a hiring decision, can anyone explain why? If we don’t fix this, AI could reinforce systemic biases without anyone realizing it. And the more complex these models become, the harder it gets to understand their decisions.”
Madan sees privacy as another major challenge. “There’s a fine line between AI-driven personalization and surveillance. Companies will be tempted to collect more data because AI thrives on it, but that also increases the risk of misuse. Self-regulation won’t be enough. We’re going to need strong legal frameworks, and companies that think ahead on this will have a competitive advantage in the long run.”
The Future of Engineering Careers
What should software engineers be doing to stay relevant? Madan’s advice is clear:
- Master debugging and system-level thinking. AI will write a lot of code, but it won’t always be correct. Engineers who can diagnose and fix problems will be in high demand.
- Develop a deep understanding of business and product strategy. The best engineers will be the ones who understand why they’re building something, not just how to implement it.
- Invest in leadership skills. AI will handle execution, but humans will still need to drive alignment, collaboration, and high-level decision-making.
“The engineers of the future won’t be defined by how many lines of code they can write,” Madan says. “They’ll be defined by how well they can shape technology to serve the business. The best ones will be indispensable, not because they can code, but because they know what to build and why.”
Looking Ahead: What’s Next for AI?
So, where does AI go from here? Madan sees a few key trends:
- AI-powered personalization will become standard—but companies will need to balance it with privacy concerns and transparency.
- Low-code and no-code tools will evolve—allowing non-programmers to build high-quality applications, which means traditional engineering teams will need to differentiate in other ways.
- Debugging AI-generated code will become a core skill—since AI will make mistakes, but fewer engineers will have the expertise to fix them.
The best engineers will be those who can work across disciplines—combining technical expertise with business, product, and leadership skills.
“AI isn’t replacing engineers,” he says. “It’s changing what engineering is. The best thing you can do is adapt and position yourself at the decision-making table. The ones who succeed in this new world won’t just be the ones who write the best code—they’ll be the ones who define what gets built in the first place.”
For indie hackers, founders, and engineers, the message is clear: AI is a tool, not a threat. The real question is, how will you use it?
