As title states. Just curious ballpark number how many I would need roughly
If anyone has experience/anecdotes with getting interviews from OSS contributions I would love to hear. Any specific on what specific things you did/worked on/things that impressed HMs/recruiters would be appreciated!
First off, the quality is going to be more important than the quantity. Quality actually cuts across 3 different angles:
Of course, quantity matters as well. Submitting 1 beautiful PR for a high-impact issue in React won't land you an interview at Meta. In terms of quantity, it's hard to say as it will matter a lot based on context (which repo it is, the company behind it, how much they're hiring, your seniority, etc), but if I had to put out a number, I would say 10 at the absolute minimum, probably closer to 25+ to have a real shot (the market's rough right now too). I wouldn't be surprised if you need 50+ merged in PRs to start seriously playing ball.
Similar to side projects, open-source is a long-term endeavor. There's no such thing as a couple week thing you can do to give yourself a differentiating edge in today's job market. All of these things will take 2-3 months minimum. For big, long-term gains, you need to put in big, long-term efforts.
If you haven't already, I highly recommend checking out our open-source course here (made by the former Director of Open Source at Meta): Become An Open Source Master
I definitely agree/understand your point
(1) I'm curious to understand how much quality affects getting interviews.
Assuming the HM takes time to go through your PRs and assess the quality, I can see it making a difference. But I am not sure how many HMs actually look at the PRs being submitted
My hope is that I have heard anecdotally from friends that after sending a few (non trivial) PRs to brand name repos e.g. meta/MSFT/google the number of callbacks they got increased. But this could just be survivorship bias. I'm definitely willing to spend a few months on this
(2) I'm curious to hear anecdotally your experience with screening candidates with Open Source contributions, how often would you examine their PRs and verify quality?
I'm curious to understand how much quality affects getting interviews.
I am primarily talking about the case where you're looking for an interview from the company behind the repo - The dream scenario is that the maintainers of the repo observe your contributions over time and eventually give you a glowing referral into their interview loop.
As for companies judging outside open-source contributions, I'm unsure how those are weighed. I feel like if a hiring manager is nerdy enough to care about open-source though, they will probably ask for a reference to confirm the quality angle (instead of manually checking themselves, which is effectively impossible).
I'm curious to hear anecdotally your experience with screening candidates with Open Source contributions, how often would you examine their PRs and verify quality?
Unfortunately, I have literally never had a resume come through with meaningful open-source contributions. This is a combination of 3 factors:
I wrote some of the core code for LlamaIndexTS in Fall 2023. Volume of code wasn't very high, codebase was small. Someone was willing to take me under his wing and provide some guidance on navigating open-source. LlamaIndex was really small at the time and they completely blew up in popularity. I then made a Linkedin post that generated 10+ interviews within a week. I was so overwhelmed that I declined 8 of them.
Fixed the resume and got 2-3 interviews a week online via cold applying. My Leetcode skills were horrible, so I literally got steamrolled by DSA interviews for 3 months. Around March 2024, my interview machine kind of stalled. So I just started re-responding to the founders who reached out from my Linkedin post. In April, some founder on Linkedin said they were hiring and I DMed them. Signed the offer a week later and started the day after Mother's day.
Concurrently, I reviewed a book on LlamaIndex and I got 2 paragraphs of fame for free (with some part-time review work of course). My only regret is not taking the time to shout out my entire extended family.
My timing was perfect, almost nobody else was doing GenAI open-source, and the code quality bar was very low. Open-source is getting harder. It still works, but as with any market, any new alpha is arbitraged out eventually.
That being said, exceptional execution is always rewarded handsomely. I think the open-source courses here are a great head start. You don't have to wait for a Jedi master like I did.
**It'll probably take 3-6 months to get any real momentum. 1-2 if you seriously put an aggressive full-time effort into it.