Code Quality Q&A and Videos

I am a machine learning engineer and switched my job 3 months ago. My previous company did not have rigorous code review process, owe it to culture or lack of care for good code quality. In my current company, I am doing decent job but at times I really am impacted with code review process. I sometimes have fear that I'll lose job if code quality is not up to the mark, I get some good feedback from team lead for code quality - which I think is good for team as well as my own learning. In the last week, I got a few nitpick reviews - it was a PR with lot of code and I would ideally want to reduce the number of these nitpick comments in the near future.

Any suggestions on how to overcome the feeling of guilt, fear and at the same time improve the code quality? How do developers deal with it?

Additionally, there have been a couple of time when I was caught off guard - I did not know the reason behind choosing a particular approach completely. I sometimes freeze when having such conversations thinking what would the person in front of me think about me if I reasoning is not correct. Any suggestions on overcoming such scenarios in a nicer way?

I’ve been asked to write up some docs on coding standards and code review on behalf of a dev team that is stuck in its ways.

How could I facilitate that conversation?

How would I structure these docs as I seem to need to write a mix of documentation that includes high level goals with quite specific process docs

The CTO's perception is that code quality is low, but can’t or won’t articulate why he thinks that.

He also thinks the code review process is sloppy but again won’t articulate what has happened that has led him to think that.

he’s also said don’t go all war and peace on this

it’s all so vague.

Hi everyone, I recently started as an Engineer at Meta, and my team has a lot of frequent broken, flaky tests. Like every single week a test must break. The End to End Test Wiki is not enough for this. How can I improve the testing system? Are there resources that you recommend? What suggestions do you have for me?

Thank you!

Background

I have joined a new company recently and was tasked with building a Task Management System ( for clients and internal Employees ). So in short this software is going to be re-iterated again and again for improvement after we launch its first MVP ( targeted 2 months at max ).

Scenario

I have come across a use case where I want to handle errors using a custom error handler, its a function to which I pass any error thrown my APIs and based on error type ( error code a reference from prisma docs ) it returns a specific structured error response that can be useful for frontend devs and also help debug better.

Question

Is going this far for handling many errors worth it, i was recently listing to Alex for the Code Quality Course and what i understood its worth it as it would help me and my juniors debug better while in production and local development as well.

Tech Stack

Prisma ORM, Postgres, NextJS

Hi,

I'm a solo fullstack intern at a company right now; no real mentorship and guidance (no code reviews etc) so I'm wondering what steps I could take to way to grow as dev here without a real mentor or team. Feels like a glorified side project

Trying my best to learn as much as I can and grow as an engineer while building out the features I have been assigned. Not 100% sure i will secure a return offer (i probably will take it for now), but going through the other taro courses on resume + behavioral to prep

Secondly, how would one track metrics if I don't really have any? Would like to include some numbers on my resume for the work I've done at this role

Thank you!

I'm a software engineer currently working on a dynamic page that replaces a more static, hardcoded version. The issue is that much of the existing code was written with a "just make it work" mindset — hardcoded values, minimal abstraction, and very few edge cases handled.

Now that I'm trying to do things properly and account for various scenarios, I'm running into situations that weren’t considered before. When I raise these edge cases or try to clarify requirements, I feel like I'm coming across as scattered or overly concerned with "what-ifs." It’s starting to feel like I’m that person who overcomplicates things, when in reality I’m just trying to build something maintainable and future-proof.

Has anyone dealt with a similar situation? How do you raise valid concerns and push for better practices without seeming like you're overthinking or slowing things down?

I am working in a project where we have a pipeline which runs automated tests, lint, and other type checks.

But we are merging PRs even if the pipeline fails 😅

On my case, every time I got a pipeline error, I fix it in my PRs, and some of my coworkers are starting doing the same, but still we are merging some PRs with the pipeline failing.

Our manager is a software engineer too, and has the role of merging the PRs.

I tried to convince him to avoid merging PRs if the pipeline is failing, but while he is open to discuss this topic, he thinks that since other teams also need to merge things. He doesn't want to block them because of the pipeline.

More context:

• It is a startup and we want to get the job done faster.
• I have 2 months working there.
• The pipeline was always failing because of another step that was removed recently. I think they got used to ignore the pipeline because of that.

I believe we are paying 10x of the future time, for short term quick time (10 minutes of the future for 1 minute today).

I thing If we continue with this, all will blow up in our faces.

I am tiring of fix the pipeline almost every day, and checking my team PRs as well.

Not sure if I should just keep pushing, or stop worrying if the pipeline passes or not and just see how the things blow up, and then try to convince the team of keep the pipeline passing as a strict requisite to merge a PR.

What would you do in my case?

Background

I have started developing an internal task management system based on server actions. The problem is that server actions cannot run in parallel. At some screens using Parallel routes, I am calling more than one server action, and since they run sequentially, it’s hurting performance.

Question

What would be the best course of action here, I found a library that can be used to run server action in parallel While I was reading somewhere that one should not use server actions for GET requests and only for mutations, based on this I am thinking about going to ‘api’ based routes for GET requests and keep the post requests as server actions. Need advice from seniors here who have worked in NextJs before.

Also having a codebase following two kinds of approaches server action and API routes seems weird to me because as per my limited knowledge plus what Alex taught in his , one should follow one pattern in a codebase so that it stays consistent. Would appreciate the help or any kind of advice here.

a lot of my work depends on writing pandas/ETL pipelines in Python

pain point #1: I find that my functions (each logic block is a function) tends to be hardcoded. How to avoid this?

Example:

def agg(df):

    df['A'] = df['B'].groupby(['C']).apply(function)
    # 10 more lines of specific business logic

    return df

def remove_duplicates(df):

     # gets a list of existing IDs from database and removes them from this ETL pipeline

     existing_rows = get_existing_ids_from_db()

     df = df[~df['id'].isin(existing_rows)]

     return df
def get_existing_ids_from_db():
    return set(db.query(Data.id).all())

As you can see the issue is that a lot of these functions tend to be hardcoded. Thus its hard to

1. Isolate and verify
2. Extend
3. reuse

But when I make it less hardcoded, it ends up being more redundant and thus hard to maintain because if i change the logic in 1 place I need to change elsewhere (eg, remove all duplicates except this super admin) for pipelines A,B,C but not D

This is a simplified example, but illustrates a larger point that my gut instinct is to write a pipeline, then isolate logic into functions, then I realize damn now my logic is all hardcoded when I want to reuse this with slightly different cols/logic

pain point #2: when I pass in data frames into a function it makes the code hard to maintain as now bc this function expects a very specific dataframe with a specific set of columns. How should I be thinking about designing functions here

Additional ask, any suggestions on resources for good code for ETL/data engineering?