Debugging Q&A and Videos

I am a mid-level machine learning engineer in the advertising field, and I often need to provide PMs with performance numbers that are not available in the dashboard so that they can communicate with clients or make business decisions. Sometimes this need is urgent, such as needing to write code to find the number of customers who made purchases in the past seven days that meet a certain condition within half a day to a day.

I have encountered a few problems:

1. The conditions are different each time, and the code is difficult to reuse. I am not sure if it is worth investing time to write unit tests. Is there any way to make sure my code is bug-free?
2. Sometimes the calculated numbers do not seem intuitive, and I am worried about providing wrong numbers, so I spend extra time checking them, but I am also told that it takes so long to calculate such a simple thing. Indeed sometimes the code itself is not difficult to write, but the main time is spent verifying the correctness of the results. How to communicate this?
3. Sometimes I ask my colleagues to help me check my code or results, but I am worried that my manager will think that I cannot complete my work independently and that I am not meeting the company's expectations of a mid-level engineer.

These are tremendously tricky to deal with, so what are your best strategies to navigate these pesky bugs?

Under the assumption: it only occurs in the highest level prod environment.

On most projects that I worked it was enough to have a high-level of how the higher picture of the code works and for the most tasks it was enough to copy paste and adapt existing code and for debugging it was also enough to have grasp the high level image.

However, I worked only on maintenance projects, that weren't very innovative. In order to have the chance to work on a good project how well should I know the tech stack?

For example for the stack: java/spring/elasticsearch/kafka/AWS, should I know the newest annotations or how the framework works under the hood, and probably be able to extend the framework, how to debug machine on VM etc?

Is there any sort of metric with respect to how in depth you should know your tech stack? Because in order to write code in a tech stack you don't need a lot of time to learn if you know the basics, but to write state of the art code I feel like you need years.

When does being a code machine act as a disadvantage?

I identify as ND (NeuroDivergent) and when I am able to build an app in a day (like hackathon style), fix a bug, solve a problem, the disadvantage I have noticed from time-to-time if I am not highly cognizant of my ability to take breaks (timing my tasks), I forget to eat meals or do basic things (like use the bathroom)? It is much like the stereotype of Poppy Li in Mythic Quest unfortunately. How do we manage this better when there is a value in being productive in output, but an undervaluing of our bodies, and sustainability?

Do we need extra reminders on our Apple Watch (or other Smart Watches), pomodoro technique to be able to tell us to take more breaks even if we are on a roll finishing a ton of code? What strategies have worked well for the code machine archetype to get their work done but without burning themselves out?

I am currently doing some ETL (extract transform load)

But the Transform part of the operation takes a minute or two. I am having issues with the "load" part of my script where I need to update a DB with the transformed data

Pain point: It takes me a minute or two to get the transform outputs and it really slows down my coding velocity + feedback loop bc most of my time is spent waiting for the code to run

So, I'm the only frontend developer on a mobile application. My boss is BE and so if I ask for help he just tells me, "im sorry but I have my own things, you need to figure this out". I've expressed concerns when I wasn't happy with that answer; but, he doubled down that I knew more about him than my problem and so couldn't help me.

The issue is, the things I have problems with are exactly specific to frontend, maybe I'm trying to do some data flow stuff and just want to bounce off a coworker. Or, I have to integrate some FE piece to the BE and since we're a 3 person (engineer-wise) startup, we don't have documentation or really anything besides slack messages to explain stuff.

This has led to me being forced to just white knuckle my way through problems. For the past year and a half I've been able to do this; however, I'm now facing more difficult problems, live-streaming, bridging native modules ( I work with RN ).

More recently, I got stuck on a problem where, I seriously contemplated quitting the company because I couldn't figure it out. There is a ton of pressure because we have daily stand up and I can only say, "I'm still working on X due to Y" for so long. And so I thought, what happens when breaking it down, trying to solve a simpler problem, posting online, talking to teammates, reading docs, just doesn't work? I seriously thought everything was spiraling out of control.

I honestly don't know if there is an answer to this problem. But I was truly feeling hopeless just blindly trying to solve an issue by googling, chatgpt, and hoping for the best each time I hit compile.

Code freeze is coming up next week so everyone is scrambling to ship their diffs in right now. However I am facing a hard technical blocker which can be resolved by a coworker on another team, but he’s too busy with his own diffs to help out. I am unable to reproduce or debug this issue myself as it involves some highly technical details about how a proprietary mobile framework (Bloks) interacts and compiles down to the Android native fragments, and I have minimal experience with either. The coworker wrote that specific part of the Bloks framework and no one else seems to understand it (even the Bloks on-call couldn’t figure it out). What can I do at this point?

Since a week I have been investigating a production bug that last occurred on 20th Sept 2023, and earlier in 2021, but unable to reproduce it in test environment. At times there's a transaction that erroneously overrides some database values, and I need to reproduce a similar transaction to see where in the code the database value is getting overridden. I followed the following steps:

1. Got the request URL from logs for the buggy transaction, to see which UI page is responsible for it.
2. Spoke to the user who performed that transaction if he remembers what he changed on that UI page. It was a POST request and only the URL is logged, not the body. He faintly remembers what he changed. The database history tells me what columns were updated exactly for that transaction, but those columns are not visible on the UI. Those were updated indirectly as part of the page save. But I have not figured out still in what scenario those columns will get updated in the database from the given UI page.
3. Tried performing the same steps in test environment, but unable to reproduce. Tried for different scenarios by changing different fields on the UI, since the user doesn't remember what he changed exactly.
4. Will next be debugging the code on my local system to get an understanding of the flow.

I'm worried - what if I'm not able to reproduce it? I often think that others will doubt my capabilities. Can I be put on PIP for this?

A year back I joined a Big Tech company as a mid-level software engineer. I had 5 years of work experience mostly in not-so-famous startups and I joined a large tech company after doing my masters.

It's been a year since I joined but I regularly feel like I don't belong here. I go through alternating waves of confidence and self-doubt. When I am not able to debug simple issues in a new microservice, I feel dumb. I feel like the senior devs on my team are just able to solve everything and I am still struggling after a year. I have been through a round of layoffs and re-org and am not sure about the kind of work I will be doing in the future.

I want to be promoted to senior engineer level but constantly get feedback that I am not assertive, opinionated, and take more time than usual to complete ambiguous tasks. I see everyone getting promoted around me and I don't understand why I can't seem to be improving. I am very motivated and willing to slog hard, but it seems like I simply don't get it or am not smart enough. Everyone around me just feels smarter and more experienced.

I feel like moving to a smaller company with not-so-high coding standards and ditching big tech because it will be more up my forte. I am aware that I won't grow there. It just feels frustrating to be stuck at a junior-mid level, 7 years after my bachelor's. But I also know I am not at that level yet.

Any advice on how to go through this problem to the other side will be lovely.

I work in as an iOS engineer. My current company is a startup that's scaling up fast. In classic startup fashion, testing was historically not a top priority here. As a consequence, we struggle with lots of manual QAing and low code confidence when modifying pieces of poorly tested legacy code.

This has improved in the past year simply by unit testing every new piece of code by default and by introducing unit tests in legacy code as we touch it, boy scouting style.

At this point, I think we need to step it up and I'm looking forward to formalising a testing strategy for our mobile team.

Here are some ideas I have:

• In the realm of unit tests:
  • I want to introduce API contract tests because our back-end API lacks documentation. In these tests we'd unit test the way we build our requests to API endpoints, and we'd have a sample JSON response from the server to test that the app can parse it into our data models as expected.
  • I want to introduce localisation tests because our app supports many languages but we are always manually testing localisation. In these tests, we'd assert that the outputs of our formatting logic are the ones we want for different locales.
• In the realm of user interface tests:
  • I want to introduce snapshot testing, because right now the only way we can test that our views are rendered as expected for different view configurations (for example RTL vs LTR, empty content, loading content, content in verbose languages, etc.) is manually.
  • I want to introduce UI tests (sparingly, only for key flows) because we have no e2e tests. Currently, key flows are tested manually and frequently to prevent regressions, which takes lots of time.

What other testing strategies can I propose to establish some standard we can use to further improve our code confidence, reduce bugs, and rely more on automated testing?

How do the top tier tech companies approach mobile testing?

I have just joined a new team, and have my first coding task. I am currently terrified about taking the first step and writing some code and testing it.

I have spent a lot of time researching solutions, but have found myself bit in the past because of a lack of testing and research, and also am just terrified of breaking things.

What are some suggestions for getting over this fear, and how can I be more confident that the testing and implementation cover all edge cases and won't have bugs?

I'm on-call this week and I'm met once again with a strong sense of not having a clue what I'm doing. I know the majority of the job of engineering is trying to work within a legacy system but I feel like I'm missing tactics to help me make any progress at all. That and having to switch contexts to address alarms and queries has meant that I actually made no progress today whatsoever so this is part rant and part question - any tips on confronting a wall of overwhelm and making progress towards the most impactful bugs?

As part of my story or task, often times I have to understand and contribute to other codebases or solve bugs that involve understanding flow of other codebases. What is the best way to navigate this when you are suppose to solve an issue where you have to understand an external codebase in order to contribute or fix an issue?

Every time, when my manager asked me to do some changes to the repository that is totally new to me. I became scared.

I prefer to do research by myself first. But I got lost in the new repo by reading file by file, and don't get the clarity.

So I ask the repository owner to provide documentation, mostly they don't maintain documentation, and even if they do, it is not updated or it involves a lot of detailed feature-wise documentation, which is usually not relevant to my requirement.

Then, I call the POC of that repo, but I couldn't figure out what is the right question to ask in the first call. Over time, I ping him asking questions whenever I face hurdles while achieving the requirements.

Sometimes, I put a debugger or logs to understand the flow of code.

The above processes took a lot of my time.

What is your suggestion to get clarity in the new repo such that I can complete my requirements in less time?