HN Explorer

Ask HN: How should MLE navigate career in today's industry?

1 points by throwaway123198 ↗ HN
As an MLE how are you navigating your career right now? AI Labs want researchers to work on the modelling part. The rest is more infra style software engineering.

I feel that the rest of the current set of applications is either being replaced with calling LLM APIs or is so specialized that generalist MLEs aren't chosen anyway.

RecSys is a standout for me. Mostly being done by the largest platforms in the world but also lots of challenges ahead of those businesses, especially as the web becomes more agentic and designed for agents overtime. Also, the teams tend to be heavily dominated by single ethnicity teams. Is it worth it to invest career there?

Or is it better at this point to try to become an "AI Engineer" or just leave the space altogether?

2 comments

[ 0.24 ms ] story [ 9.7 ms ] thread
packageman
Ai make bigger mistakes possible. See Zillow/iBuy nuking $29B market cap & 50% employees eventually lost their jobs because they ignored direct warnings that they were poisoning their own algorithms. ML engineers that cultivate executive communication channels are, IMHO, invaluable.

Want another example of how MLE's matter? A satellite image machine vision algorithm designed to 'see' soviet tanks used photos without tanks on a sunny day and with tanks on a cloudy day. Guess what - the algorithm only worked when it was cloudy. This was worse than the husky/wolf dilemma. It was just an expensive weather detector and an MLE with an organizational voice could have saved millions of wasted tax dollars.

Happy to provide Zillow documentation (the Soviet tank example is old, unclassified, and easily findable).