HN Explorer

Ask HN: Early 30s SWE feeling stuck

1 points by rafiki6 ↗ HN
I'm an early 30s SWE with a MSc Data Science recently completed. Have about a decade of diversified experience, but mostly app dev in financial contexts. Currently working as a Senior at a financial institution, but little to no growth left for me here as I am trying to get much much more ML and Data Science exposure, and all such projects have been shut down here since pandemic.

I have gotten interviews at other places, but everyone has adopted leetcode style at this point and I haven't managed to get through. Since interviewing this year, I've probably done about 7-10 interviews, and have only made it past first round a couple of time.

I keep practicing leetcode and trying to get my DSA chops back to where they need to be. I have started exploring ML and DS side projects to keep my self fresh in that space, but I'm worried I am getting left further and further behind my peers and at my age I only really have ~5-7 years or so before ageism sets in and I need to leave pure IC roles.

Any advice on what to do at this point?

3 comments

[ 3.3 ms ] story [ 17.1 ms ] thread
ianceicys
Read the book, the power of habit. Here's a few other books I recommend:

1. The Unicorn Project 2. Measure What Matters 3. Start With Why 4. Project to Product 5. Accelerate

brutus1213
I shifted to doing AI (mostly vision) stuff a few years back, and I sometimes feel I was happier doing backend systems. There is so much noise out there in AI and the whole thing is moving at the pace of JS frameworks (i.e. too fast to humanly stay up-to-date with). I don't know if it is generally better to make the switch to AI or stick with the old stuff. People will always need the old stuff. Also, there is a reason those AI projects have gotten canceled at your workplace btw .. I think most orgs don't have the right setup to derive value from AI.

Question to OP if you want to discuss this .. how do you see ML/DS vs. vision. Vision needs a ton of data and I think hard for a lot of orgs to get benefit from. Is ML/DS very different?

rafiki6
Agreed on most of what you said. DS doesn't necessarily need a lot of data. Alot of DS work is not much more than visualizations and report generation to find insights.

ML might need a lot of data depending on how your org defines it. There are many interesting non-deep learning applications out there.

At my org I was mostly involved in NLP work. Honestly I think that will be the big game changer for most of the Enterprise, as they primarily deal with text data. Think automated parsing and analysis of emails, chats, reports etc. Something like GPT-3, can be a game changer if/when it doesn't need a multi billion dollar super computer.