How deep can you go in your past experiences?
Posted by proof_required@reddit | ExperiencedDevs | View on Reddit | 13 comments
I have more than 10+ years of experience in ML/DS and my resume has relevant experiences from 10 years ago. One time in interview some younger kid was making fun that I was using SVM. I didn't get offended and found it funny myself how things change so fast.
I know that once you put something on your resume, it's fair game and you can be asked about it. But I sometimes struggle with the grilling I get on things I worked on a long time ago, especially when I was still pretty junior and not really the one making the decisions.
How do you handle questions like, "Why did you do it this way?" or "Why didn't you choose a different approach?" when the honest answer is that you weren't the decision-maker? Other than making things up, what's the best way to answer those kinds of questions?
tyr--@reddit
Well, I think it depends a lot on the context and what the entry in your resume or mentioning a past experience is supposed to mean in the context of the role you're interviewing for. With 10+ years of experience, why would you mention situations where you weren't the decision-maker, especially if you're aiming for a senior/staff-level position? Just to show breadth of knowledge or is there another motivation?
proof_required@reddit (OP)
Yeah mostly to show the breadth of knowledge. How far do you go back in your resume?
tyr--@reddit
I work a lot in high-performance parallel computing and there's work I did almost 15 years ago that is sometimes relevant (some MPICH implementations), but I mention it from the aspect of how I used that tool that was available to me to solve a specific problem, not to just say I'm familiar with MPI.
Same for early machine translation work (think pre-transformer and LLM times), I'll mention I worked with statistical MT and systems like the Moses decoder, but again not because they're relevant to the job today but because some of the optimization mechanisms and approaches can be reused.
hkisthebest@reddit
For my learning purpose, how would you phrase that experience in a bullet point?
kbielefe@reddit
Don't ever just say "it wasn't my decision" if you're applying for a decider job. Say something like, "It wasn't my decision, and if starting fresh today I would consider using instead, but it has benefits in ..."
They don't really care why you used something 10 years ago. They really want to know how you make similar decisions now.
Bubbly-Watch6214@reddit
When I interview, one of my favourite answers is when people just straight up tell me they weren’t the decision maker. We all start somewhere and for most of us, that’s with a boss. If you have that amount of respect for your career, you’re already doing well. If you can take that experience and show how it helped create your current taste for software, you’re a favoured candidate.
beneath_the_knees@reddit
Thing is, even in this day and age where neural nets are all the rage, there are still many scenarios where old fashioned classic models out perform them. Hell, sometimes even bob standard linear regression can out-perform them. So next time just tell them to STFU
Realistic_Yogurt1902@reddit
I guess it depends on the companies you are interviewing with. I've had several dozen interviews over the last 10+ years and was never asked about anything super specific from my CV.
proof_required@reddit (OP)
Yeah it's not super common. Happened like once or twice since the interviewers weren't really familiar with my recent work and found those experiences more relevant.
Adept_Carpet@reddit
Generally with older experience I isolate the part that is still relevant.
Like, experience with the CodeIgniter framework and the YUI library can turn into experience with PHP and JS. The stuff I made in Perl, I just call it a web application.
With SVM, maybe call it a classifier. If they ask for why, SVM is easy. It was probably the state of the art model for the problem at the time, at least if you considered the tradeoffs involved.
CaffeinatedT@reddit
- "T'was was the fashion at the time"
- Committing and adapting a working system to work as well as it nearly always beats starting from 0 on a marginally better thing that does the same thing, see also: Rewrite fallacy.
LethalBacon@reddit
I just literally say it wasn't my choice, but try to give reasons for why the decision makers chose that specific route. And typically, the answer is that it was time/technical debt related.
We used x pattern because some unknown person determined that architecture a decade+ ago and we didn't have time/resources/budget for a full rewrite, so I had to work within the constraints given to me.
darkhorsehance@reddit
Tell the truth? If it wasn’t your decision, tell them it wasn’t your decision. If it was your decision, explain why, including tradeoffs.