Comment by goldemerald | Hacker News Reader

Comment by goldemerald | Hacker News Reader

I am going to write a lament about being a machine learning PhD (near) graduate. Whenever I see these very successful new PhD's, I just have an overwhelming sense of unfairness and sadness. Many of my papers were borderline rejects, delaying me months every time until there's no way for me to catch up with the likes of the author. I am attending a top 25 US school for machine learning (but not better than 25), and the magnitude of difference for the quality of PhD students is staggering. Rarely does anyone at my school graduate with 3 first author papers, let alone 5 (with so many orals!!). It consistently blows my mind how much better these top 10 schools are then the rest at producing papers: the only metric that matters. I've thought about applying to OpenAI as an engineer, which to me would be an amazing opportunity and very prestigious, but I don't want to waste my time: I only have 2 papers done (and God willing a 3rd one for CVPR in 3 weeks). Naturally the skill ratio between me and the author is at least 5:2 (probably worse, since I don't have any oral papers), but this doesn't stop me from feeling that I could contribute to OpenAI as effectively as any new PhD. I know if I were a hiring manager, I'd always go with the more proven candidate, but as the applicant I wish I could prove that my lack of publications should not be held against me. Perhaps this is a common occurrence for any engineer, hitting a ceiling for jobs in this manner, but that doesn't stop me from wishing differently.

Look, I'm from a university you surely don't know. Not in the US, and not even top 500 in the world by most rankings.

Lately I had two PhD students: one graduated with 4 papers in undisputably top-tier conferences (plus 2 in close-to-top-tier ones). The other, only with 1.

But you know what? The second student was better than the first (IMO, of course). The first got lucky hitting a research topic that proved to work very well and produced many results with relatively little effort once the basic framework was in place. The second explored more diverse research topics, and did exceptionally deep, high-quality research, but got bitten by reviewers (partly because the research had less "flashy" goals, was more niche and harder to understand, partly because of the massive random factors inherent to reviewing).

Both found jobs where they are satisfied AFAIK. The second student may not have had the best publication record, but from what I know (from being asked to give references, etc.) they found places where they went beyond mere counting and appreciated the quality of their research when making hiring decisions.

Neither university prestige, nor number of papers, are as important (or indicative of real skill) as you think (although of course they help). The world is much more complex than that.

Sounds pretty defeatist! At least try? So what if you aren't the best ever?

Why fight?

Answering that question may help you cope with not being the best at everything.. for example, I fight for my kids and employees and family. Not about me, it's about them. So what if I'm a mediocre programmer/person? To a few people I'm cool, and that's enough for me.

It's strange, choosing to stick out a Ph.D. (and not just get a masters and be happy) really cemented the competitiveness in my mind for years. I knew from the start I'd never make it to Deep Mind, but I had figured my experience as a SWE and other internships would make me a competitive Research Engineer. Now that I'm applying for jobs and not doing as well as I expected, it's a huge emotional negative for me. I tell myself "Journey before destination" and that I really can focus on my family and enjoy my life, but it doesn't stop the hyper-rational part of myself from saying "but you could have been a self-actualized top research engineer and a family man." To answer you question of "Why fight?", it's that I cannot stop myself, despite trying.

Why put your happiness in other people's hands? Why can't you do the research you want to, already?

Can you rephrase your "why fight" answer to be for anyone else? Do you consider yourself narcissistic?

Thanks for sharing your heart, will probably help someone.

Edit: another thought is to picture yourself as this mega-nerd at the top of the food chain. How do you think they feel? Do they feel inadequate? Maybe they helped make ChatGPT in a big way, but have very little equity. Maybe they are short or balding or have a speech impediment.

I promise you if you live outside yourself and for other people, it will change you. I'm certain you can help thousands underneath you...

"Why put your happiness in other people's hands?"

I ask myself that every time I've submitted a paper. It probably doesn't help that GPA is an easy surrogate for self-worth, which smoothly transformed into publishing for a self-worth measure. And the fact that I don't have particular research I want to do (just that I want to do research), is something I'll have to consider more-- probably a sign of bad foundational motives.

To rephrase the answer, I'll give some extra context. I was a SWE for 2 years before my PhD, and I was financially stable and content. The choice to give up that salary for science was not hard, but it was significant. Now that I've reached the end of my current science road, it would feel like giving up to go be a MLE/SWE. Obviously I could do a post-doc somewhere, but that kicks the can down the road again. I'm pretty sure I'm not a narcissist... I'll ask my friends hah.

Thanks for sharing.

I bet if you think a bit, you're very capable of an _insane_ amount of good. Imagine a 12 year old in your town whose single mom can't afford a laptop. Maybe he or she, this 12 year old, is more capable than even you.

You find that kid somehow, maybe they're your neighbor. Maybe with a gentle push like an old book and your old laptop... you create a great thing.

Maybe it only takes a compliment. Some dumb freshman you're TAing a class for has a rough time. You reach out, buy them lunch, talk about your struggles and success.. who knows, could really fire them up.

I bet the universe will show you a few opportunities in the next month. Don't ignore your purpose that's right in front of you, not to say you are ignoring it! Just trying to be encouraging.

I won't pretend to know how top tier ML companies hire for research positions, but for literally any other job, the cost of applying is trivial. Throw in some outbound to members on the team for a potential connection and it seems like it's worth the cost for a dream job?

Add to the fact that OpenAI is almost certainly going to be expanding wildly given that they have taken on a 10B investment, it seems like as good as a time as any.

I've never cold connected to anyone. I wouldn't know how to go about doing it. My first idea is to find members of teams that cited one of my papers, but that makes me sound neurotic about papers, doesn't it. I've had the application page for OpenAI's position open for a couple weeks now, I'll give it a serious try tomorrow, and see if there's any team member that at least 2 degrees of separation from me.

Find them on LinkedIn and ask to grab a cup of coffee or jump on a Zoom. Be open about the fact that you're interested in what makes a good member of the team. Ask about what you can be upskilling on. Be friendly.

Also doesn't hurt to reach out to potential managers as they'll be able to provide more direct feedback.

My grandma had a very funny book with Jewish jokes and wisdom (she was not a Jew, such books were popular in Poland for their unmatched sense of humor - if you like Zucker Brothers movies you will know what I mean).

Somehow the book survived the Second World War and Warsaw Uprising. As a kid, when I was visiting grandma I loved reading this book.

One of the jokes goes like this: Icek (short for Izaak) was praying every night to God to win on the lottery. Night after night he keep asking for this. Finally, one night God spoke to Icek - "Icek, please, please give me a chance, buy a lottery tikcet!" God said.

So, good luck with your application, you'll make it!

FWIW, I agree with you that # of papers is not going to matter to some hiring managers. Are you strong at coding? When I've hired research engineers, the most important things I've looked for are performance in a take-home coding challenge, and their technical creativity in approaching daunting problems.

Do you have a personal website with a C.V.?

For some extra context, I have been applying every ML research engineer/scientist job that would take a new grad.

I don't have any sense for what a hiring manager wants, it's especially hard with all the layoffs, I'm sure plenty of people are applying for the same research engineer positions. I certainly consider myself to be strong at coding. I've done a couple coding challenges on hackerrank that I do well on. But I have not been lucky at getting a phone interview yet. (I do have a personal website and CV)

Sounds like you need to network more, get the lay of the land of what’s being expected. Remind yourself that while you’re not at the vanguard you’re still cream of the crop; phd tends to sap out social skills and hence for many of us finding the first job is the most daunting thing possibly in all our lives. Slog through and after that it’ll be fun sailing for the most part.

I once attended an hackaton, and a guy doing a Phd in ML joined our team.

It turned out he was completely unable to figure out how to run a python application with 2 dependencies on his machine.

He also didn't understand the HTTP protocol and kept insisting that for the users of our website to send us a request with some payload, they needed to sftp a file with the request, and then click on the website to process their file.

I kept telling him that since the server was mine, I wasn't going to open up ssh access to the world and a POST request had to do.

Of course he didn't know what a POST request is, and kept insisting on a loop that I had to enable sftp to upload files.

Ended up with me shouting, him doing absolutely nothing besides wasting time, and everybody seeing me as the bad guy for shouting.

In the end I stuck around long enough to deliver something and present it, and left immediately after; not staying to see if we had won something.

That was the last time I attended an hackaton.

As a result, I think when hiring I would be very very careful to select someone with a Phd in ML, as they might have done no programming whatsoever in their whole life.

I interview ML engineers for a company you have heard of. I do not care how many papers you have published. You could’ve published 10 papers, I’m probably only going to ask about one. I also don’t care what your GPA is for that matter.

I do care about how well you know the particular paper I a will look into prior to the interview, especially if it’s relevant to the position. People that can explain well do well. I also care about your github, if it’s well-organized and documented.

Thank you for your perspective! I could definitely talk about all my work in extreme detail, though I have quite a hard time reaching the interview stage.

I have not put much effort into my GitHub (I only have the code for 1 of my 5 projects uploaded). I'll spend some time this week updating it, maybe that will help my woes.

Send me an email (see profile), and I'll see if we can hire you. :)

Screw credentialism and paper publishing metrics.

Thank you!

you talk about volume but does quality not matter? Unless you think these top 10 Phds are beating you in both volume and quality of research, at that point lamenting is pretty unproductive.

Reply on news.ycombinator.com