Need help with a tough principal engineer
retained text
**Background** I started at my company as a junior 7 years ago and was promoted to senior a few months ago. My manager asked me to be one of the senior engineers on a new "AI-native rewrite" project: it’s an e-commerce site so it would be rewriting part of our product detail page including header and footer entirely with AI agents with close human oversight and seeing how it goes. He said I could decline, but strongly suggested I accept because of the project's visibility. I want it known too that my manager is wonderful and I have a great working relationship with him. My only hesitation was the principal engineer leading it. I've worked with him before. He's extremely smart and knowledgeable, but he leads through fear and doesn't seem to trust anyone's abilities except his own. My manager said, “just take from him the technical learnings”. I don’t normally work with him in my day to day but this project would be different. I’d be having multiple meetings a week with him. **The setup** My usual AI agent budget is $100/month. For this project it was raised to $200/month going forward. (For context, friends at other companies have unlimited or multi-thousand-dollar budgets.) In early planning, he said he wanted me running 4–5 agents at once to start, eventually scaling to 5–10. After planning and review, I started building new components and got through about 4–5 PRs in the first few days. I spent $140 in the first 8 days, mostly using a mid-tier model on high effort. **The meeting** He called me into a one-hour meeting focused entirely on why I'd used so much budget. He went through my agent sessions and asked why I wasn't reading what the agents wrote. I was reading them. He said my initial prompts were too ambiguous and I was letting the agents do too much without enough steering. That's fair feedback on its own. He then went prompt by prompt and told me to copy each one, rewrite it, and paste it into a different AI tool to compare results. The whole time he spoke in a flat, condescending tone: "Copy it." "Paste it." "READ." When he would see what I did I would get audible “sighs” an “oh dear”. It felt like being taught at an elementary-school level. He asked why I used high effort instead of medium (the default). Said it was an expectation at this point to work effectively and efficiently with AI. He said the director of engineering for AI would see my usage as a major red flag, and that if it didn't change, it would eventually lead to me being let go. Not his exact words, but that was the clear implication. He said that for the rest of the month, I’ll have the $60 left to use AI with, but after that, I’ll have to do everything manually and told me I will not be able to meet my deadline. It was so brutal that I didn't defend myself at all. I shut down and started giving one-word answers. **Where I'm at now** I've had meetings like this with him before and usually let them roll off me. Normally when we get feedback from him, it’s the whole team in a meeting so it feels less personally directed. This one felt different. This time it was just me and him and it was an ad hoc meeting. I've lost my confidence. I dread going to work and feel set up to fail. I'm hesitant to even prompt an agent now because I'm afraid of token usage. For the first time, I've genuinely thought about quitting and walking away. But the job market is rough right now, and I don't know if that's realistic. **What I'm asking** Has anyone dealt with a "know-it-all" technical leader who uses fear instead of mentorship? How did you handle it? Is my usage actually unreasonable for a project explicitly asking for 4–10 parallel agents? Or am I just bad at working with AI? How do I rebuild confidence and keep functioning on this project, or decide whether it's time to leave? Part of this is venting, but I'd really appreciate any perspective.
document fetched 2026-10-12 00:22 · canonical https://reddit.com/r/cscareerquestions/comments/1x2yrz3/need_help_with_a_tough_principal_engineer