This as-told-to essay is based on a conversation with Manoj Aggarwal, a lead software engineer at a large software company who previously worked at Microsoft, Twitter, and Stripe. He’s in his 30s and lives in California. The following has been edited for length and clarity.
I’ve worked as a software engineer for 14 years at companies including Microsoft, Twitter, and Stripe — and I’ve seen how the job has changed.
I think that in some ways, it’s a fun time to be a software engineer. There are so many tools available that you can pretty much build anything from scratch without your own skills being the bottleneck.
It’s also a challenging time for many engineers. AI tools have made coding easier, but engineers — not AI — are ultimately accountable for the code that gets shipped. Even as AI speeds up many parts of the code review process and makes it more targeted, the added responsibility of reviewing AI-generated code can contribute to burnout.
There are also concerns about AI replacing engineers. If 10 engineers were needed to ship a project two years ago, maybe only five are needed now.
I think junior engineers could be affected the most. During my job search last year, there seemed to be more openings for senior engineering roles than for junior ones, and I think companies will continue to rely on experienced engineers working with AI tools.
That’s one reason I’m making sure I stay up to date with those tools.
I keep up with AI after hours — without sacrificing work-life balance
My current job is an engineering role at a large software company, and AI has completely changed my workflow over the past year. My coding time has essentially been reduced to writing prompts for AI tools, and in some cases, I’ve been working at a pace that feels 10 times faster.
I get to learn a lot about AI on the job. Still, I think it’s important to spend a couple of hours a week outside work experimenting with AI to stay up to date with the industry. That way, I also get to work on cool projects that wouldn’t be possible to create during work hours because my day gets so busy with office work.
Taking time to experiment with AI outside work hasn’t had much impact on my work-life balance. A big reason is that I have a toddler at home, so I can’t even think about opening my laptop while she’s awake. When I do work on personal projects or read about AI, it’s typically at night after she goes to sleep or on weekends — again, when she’s asleep.
I spend under $100 a month on AI tools and built my own chatbot
Outside work, I often use Claude Code, Microsoft Copilot, and Lovable. I spend about $50 to $60 a month on Claude Code, depending on how many tokens I use and the subscription rate. I recently tried Anthropic’s Fable model for a personal project, and wow, is it fast.
One of the side projects I built is an open-source tax and financial assistant. It lets users connect AI models like Gemini or ChatGPT and upload financial documents like credit card statements, investment portfolios, and tax forms. The tool can analyze those documents and point out things like overspending in certain categories, expensive subscriptions, or whether an investment portfolio may be too aggressive or too conservative.
I used the tool to help prepare my 2025 taxes, and it helped me verify that my CPA had done everything correctly. I built it to run locally on my computer because I didn’t want to upload my personal financial information to AI models over the internet.
Read more about people who’ve found themselves at a corporate crossroads
Engineers should lean into the domain knowledge AI won’t have
One of the biggest challenges for engineers is adapting to a job that can change rapidly because of technological shifts and other factors outside your control. That was true even before the AI boom.
In recent years, generative AI has raised new questions about engineers’ job security. It’s important to keep up with AI, but I also think AI doesn’t have the high-level picture of what a system should look like or why it’s designed a certain way. That’s where I think engineers still add value. I’d suggest leaning into your experience — both your engineering skills and the institutional knowledge you’ve built if you’ve been at a company for a while.
While fewer engineers may be needed for some projects going forward, I think the responsibilities of the engineers who remain will be very important. AI also makes mistakes, so there has to be a human in the loop.
My advice for other engineers would be not to shy away from keeping up with AI. Regardless of how you feel about AI, if you’re compared with another engineer who’s using it, there’s no doubt that the other engineer’s productivity will be higher, so experiment with different AI tools and figure out what works best for you.
Do you have a story to share about learning AI or working in tech? Reach out to the reporter via email at [email protected], or via Signal at jzinkula.29.
