Most Engineers Have No Idea What They're Bad At. Here's How to Find Out.

Honestly, I'm just an SDE-1. Barely 2 years of experience. I'm not the most cracked engineer you'll ever meet. I don't mass-produce open source projects on weekends, I don't have a blog with 50k followers, and I definitely don't solve leetcode hards for fun.

But what I am good at is adapting. Whatever the task is, frontend, backend, infra, something I've literally never touched before, I get that shit done. I figure it out.

And for a while, that felt like enough.

The Problem with "Just Keep Learning"

Especially in the era of AI, every day feels like a fight to prove you can't be replaced. A fight to stay one step ahead. You open Twitter and someone's built an entire SaaS product in 3 hours using some new AI tool. You read a blog and suddenly you're supposed to know Rust, system design, distributed systems, and also be a prompt engineering expert. All while shipping features at work.

Everybody tells you the same thing: skill up. Learn this. Learn that.

And you actually listen. You start that system design course. You read those blogs. You watch those YouTube videos at 2x speed.

But here's the problem. It's too wide of an approach. You're trying to boil the entire ocean. You don't have months anymore to casually get good at things. It's a sprint race out there with the distance of a marathon. If you're spreading yourself across everything, you're not actually getting meaningfully better at anything.

The Feeling of "I Can Do More"

So recently, I started feeling this weird thing. It wasn't imposter syndrome. It was the opposite. I knew I had more potential. I could feel it. But I just couldn't figure out how to channel it properly.

There's just too much to learn every single day. Too many directions to go. And when you don't have clarity on where to go, you end up running in circles. Honestly, it gets overwhelming sometimes. You're working hard but you can't tell if you're actually getting better or just... busy.

I was stuck in that loop for a while.

A Conversation That Changed My Thinking

So I did what I probably should've done earlier. I talked to someone who's been through this. A lead engineer on my team.

I asked him straight up: "How do I get better? Like actually better, not just 'read more docs' better."

And he asked me one simple question that genuinely rewired how I think about growth:

"Can you quantify yourself? Can you visualize your own performance the way people quantify how a company is performing?"

That hit different.

Think about it. Every company out there has dashboards. KPIs. Metrics. Revenue trends. Churn rates. They don't just feel like they're doing well. They measure it. They know exactly which numbers are going up, which are going down, and what levers to pull.

But as engineers? We just... vibe it. "Yeah I think I'm decent at React." "I'm okay at backend stuff." "I should probably learn more about databases."

That's not a growth plan. That's a wish list.

So I Quantified Myself

That conversation left me thinking for days. And then I did one simple thing that honestly gave me more clarity about my skills than 2 years of self-reflection ever did.

I fetched every single review comment on my PRs from the last 3 months.

All of them. Every nitpick, every suggestion, every "can we do this differently?", every "N+1 query optimizations". All of it.

Then I fed it all to AI and asked it to create a personal engineering report.

What the Report Told Me

The report was... humbling. And incredibly useful.

It broke things down into clear categories. What I was good at, what I was bad at, and most importantly, what patterns I was repeating without even realizing.

It showed me my strengths. Things I didn't give myself enough credit for. Like shipping fast, taking ownership, and rarely needing hand-holding on new domains. The "get shit done" instinct was real. It just needed refinement.

But then it showed me the other side. And that's where it got interesting.

There were patterns in the review comments that I had zero awareness of. The same kind of feedback kept showing up across different PRs, different weeks, different features. And I never connected the dots because each PR felt like a fresh start.

I was making the same type of mistakes on repeat. Not big, codebase-breaking mistakes. Small, consistent ones. The kind that individually seem like nothing but collectively tell a very clear story about where your gaps are.

And until I saw them all laid out together, categorized and quantified, I genuinely had no idea.

From Report to Action Plan

Having a report is great, but it's useless without action.

So I turned those findings into an actual plan. Not a vague "I'll get better at coding" plan. A specific, targeted one. Where every single item on the list was directly tied to a real pattern that showed up in my reviews.

For every repeated mistake, I wrote down a concrete rule for myself. For every gap, I identified exactly what I needed to learn. Not an entire domain, just the specific thing I was getting wrong. For every strength, I figured out how to lean into it harder.

The beauty of this approach? It's not generic advice. It's not "learn system design" or "read Clean Code." It's hyper-specific to my actual weaknesses, based on my actual code, reviewed by my actual team. Your report will look completely different from mine. And that's the whole point.

Why This Works Better Than Generic Advice

Now let me be clear. I'm not saying you should stop learning new things. You absolutely should. Pick up that new framework, read that system design book, explore that side project. All of that matters in the long run.

But if you only do that and never look inward at what you're actually getting wrong right now, you're just stacking new knowledge on top of the same broken patterns. You'll know more stuff, sure. But you'll keep making the same mistakes at work. And those mistakes are what's actually holding you back.

Learning new things makes you a wider engineer. Fixing your weak spots makes you a sharper one. You need both. But most people only focus on the first one.

Here's the thing about generic engineering advice. It assumes everyone has the same gaps. But we don't.

Maybe you're great at writing clean components but terrible at database optimization. Maybe you write perfect types but your error handling is a mess. Maybe you're amazing at architecture but you keep missing edge cases.

You can't fix what you can't see. And you can't see it if you're not measuring it.

Fetching your PR comments and analyzing them gives you a mirror. An honest one. Not filtered through your own perception of yourself, but based on what your team, the people who actually read your code every day, is telling you.

The Takeaway

If you're an early-career engineer feeling stuck, feeling like you have more potential than what you're currently showing, try this:

Pull your PR review comments from the last 3-6 months.
Feed them to an AI (Claude, ChatGPT, whatever) and ask it to find patterns, common feedback, strengths, and weaknesses.
Be honest with the results. Don't get defensive. These are people trying to make your code better.
Build a specific action plan. Tie every action item to a real pattern from your reviews. Make it concrete enough that you can actually check it off.
Revisit in 3 months. Pull the comments again. See if the same patterns show up. If they don't, you've actually grown. Measurably.

You don't need to learn everything. You need to fix the specific things that are holding you back right now. Stop running the marathon blindfolded. Quantify yourself. Know exactly where you stand. And then get better — with precision.

It's not about being the most cracked engineer in the room. It's about being measurably better than you were last month.

And that, you can actually control.