Well, about 4h30m to the interview and I'm stressed. Honestly, this is why I hate interviews, because it seems the more I study up, the more I feel like I dont know what I'm doing, even though I built it...
Mostly, because during the tech interview, I was asked really in-depth questions and expected to talk through implementations, quoting method signatures, which isnt really something I consider a skill of mine...
Not to mention, last time was just with the Tech Lead... This time it's with the CEO, Tech Lead and a Senior Dev... So a simple 'I did it this way, because that's the right way to do it' wont cut it. I need to be able to explain why I chose it vs what alternatives could have been used. All things I dont generally give a hoot for.
So needless to say, I'm staring down a barrel of a gun and I know I'm gunna dance like a chicken and make a fool of myself. Idk about any of you, but I work on multiple systems at once and I dont much pay attention to anything in particular, more the overarching architecture and making sure things actually work, are tested and benchmarked. If the benchmarks are unsatisfactory, then I optimize, simple as that. I dont much consider alternatives until I have the time to explore them and I dont consider them till they're tested and benchmarked. Not quite the kinda thing that shines in an interview...
What makes it worse, is it's an hour... For the AFS and the pgrust port combined... They wont beat around the bush, they're gunna hammer me with hard questions that I know I'll choke on. That being said, I have an ace up my sleeve... My SASOS can finally run and render native Notepad++. If I can find a gap to show them that, it might distract from the fact that I'm scatter brained and cant explain why --net during build is the exact fix needed 2 layers down, in order to make pgrust print times accurately. I'm stressed and I'm anxious and rightfully so, because I feel like I dont know what I'm doing. It's kinda like a carpenter who learned by trade, being asked 'what's this chisel' and they just call it 'thin chisel'. I can write solid code, but dont ask me to explain it, just say thank you that it works and it's tested.
In other news, IDE got a major upgrade. I added nodes! You can now host a headless running on an ubuntu server, then from your GUI instance, allocate work to be done on that system, so if you have a few old systems lying around and you wanna boost your productivity, you can now work on 1, execute on many. Next is cloud VM scaling, where a cloud instance can trigger expansions (think Kubernetes) for each agent it spawns, so ephemeral state agents can execute in their own sandbox environment, while maintaining connection to the sitemap. That way like you have a local LLM and a cloud LLM, you can just 'offload to cloud' for heavy builds or if you really want to power through a complex system via the teams studio.
Idea for that 1 being, wasmer's edge is actually perfect for it... Think about it, if an agent spawns as a .wasm instance, you can scale thousands of agents and scale back down to 0, cuz each agent is ephemeral, they just have their skillset and objective.
See... Scatter brained, I'm meant to be stressing about the debrief and I'm talking about my side projects! It just feels to me like I'm suited for working, not for interviewing... Cuz these are the kinda things that add value on-the-job, not before you land the job. They're looking for competence and I'm giving them experimental 'I'm not quite sure's' and that's what stresses me. Relaxing didnt help much either, clearing my mind came up foggy at best and I worry that I'm gunna trip up on the finish line. It took them a day to review the work... In that day, as the people who maintain the system, they most definitely knew exactly where to dig to find the flaws and the things I never thought of. An 'I dont know' or 'well that's why you're the lead' wont work... This time the interview isnt just with the lead (he's nice, during manual code review, he actually gave me hints, cuz he wanted to help), this time, it's him + 2... I dont know the other 2... I dont know if they're entering constructive, or destructive, if their goal is to weed me out, or to see my potential. Cuz I'm pretty sure I'm not the only candidate in the running at this phase. Probably am the only 1 who clocked in less hours than their predicted minimum though... And only 1 who logged AI hours at over 100h... But those can count to my benefit, as much as they can count against me. What I see as optimization, they maybe see as lazy ai prompting if I cant explain it.
So now that I've had my existential crisis and I've accepted my fate, come what may, I've got 4h left to prepare for my impending doom... Social anxiety's a bitch aint it.
Wish me luck (I'm cooked...)
Top comments (13)
I can't believe getting a job is this hard these days 😮 Why are they taking it this seriously??? 4.5 HOURS for an interview??! What might those other candidates be doing?
EDIT: The interview is only 1 hour. My mistake.
No no, the interview is 1h, but it's at 6pm CET, so 4.5h to the interview
Oh. But still too much, honestly.
Yeah, but 1h is a safe bet for thoroughly going through things and discussing what comes next. Also given that how it goes, we'd likely discuss the offer? So fingers crossed they dont try to screw me...
I mean the overall process. The offer is the most awkward part, right? Given that they care this much, you'll probably get a solid offer! Just make sure to be calm, as xulingfeng said.
Fingers crossed! The main thing is they're a small team in a startup, I'd be person nr 10 and that means that they dont really have the capital to offer top bracket... But lets hope they dont low-ball too low...
So strange. You are stressed about the interview, but want the highest offer possible 😄
Well yes and no... I just dont wanna get screwed over.
LMAO, you’re just overthinking, stuck in anxiety and mental exhaustion.🤣 My advice: ask your fiancée to give you a heavy knock to snap you out of it, grab some KFC if you can, and look up yogic breathing. Be a man and stop being so indecisive.😏
You say this, as you know KFC is a town away 😭 You are right though. But it was just my mind looking for strings and I found them! It took me a while... But I now understand what this debrief is for and why they asked me to design both systems... Because their AFS system is exactly why porting pgrust is a mess! That's why they wanted me to design a POC and deliberately leave it lean, so after I am done with both tasks, during the debrief, we can discuss how my design would need to be expanded in order to replace theirs and how it would affect eg. pgrust!
In the end, an anxious mind, is a mind seeking answers and once it finds them, there's a calm. So I can now confidently say, I'm ready for the debrief, cuz I know exactly what they're going to ask and I know exactly why they're going to ask it too. They admitted that their AFS system is in a bad state and it shows, half the fixes to make pgrust work, were workarounds, because they didnt stick to POSIX, which is a pretty big oof, given they market WASIX as expanding WASI to add POSIX capabilities. It all makes sense now!
It never dawned on me while fixing it, because I thought it was separate tasks, 1 experimental, 1 implemental, turns out the goal is how to turn an experimental and implemental into 1 production grade codebase! Which is funny really, because POSIX AFS is exactly what I ran into with the Notepad++ implementation, because of roughly the exact same reasons... The only other issue with Wasmer's side is that they dont have durability really figured out, which I sorted out during my initial tech demo! Combined with the AFS I designed, pgrust would have just worked, not even need patching, because it would have been compliant and durable with a single command addition.
Ooohhhh, this is good, I wonder how many of the other candidates connected the dots? Cuz I only found the threads when I went outside of the scope to optimize, they werent recycling worker threads, which with a basic app wouldnt cause any issue, but with pg opening and closing connections, it'd cause exhaustion within an hour... And without a durability layer, pg would just die, it wouldnt be able to checkpoint or reload at all, just dead, spin up a new instance. So now I know exactly why they're hiring me and what I'll be working on first, cuz why assign me to something else, if I already built out 90% of what's necessary to fix their core infrastructure?
Awesome! Glad you’ve figured it all out. Just relax now. I can send you some photos of KFC here if you want, showing exclusive Chinese KFC menu items unique to China.🤣
Sure, why not, I'm kinda curious about the exclusives you get there 😂
This shit is real!