1–10 of 55 posts
It functions by matching you 1-on-1 with a compatible developer every week :)
Show HN: DevClad – A social-workspace platform for developers
github.com
Is this like Lunchclub? https://lunchclub.com/
I built an even more rough version of that here - https://connectdome.com (but then I realized I was dealing with feature-bloat) so now I'm experimenting a little slower.
I must admit I rolled my eyes when I read "ML algorithm" since it's usually marketing bullshit. Seemingly it's unsupervised knn w/ Scikit Learn, mixing variables like "open to video calls", "location", "openness to ideas" and "timezone".
Fairly neat.
> Every week, you are matched with another dev using an ML algorithm for a 1 on 1 call. I must admit I rolled my eyes when I read "ML algorithm" since it's usually marketing bullshit. Seemingly it's unsupervised knn w/ Scikit Learn, mixing variables like "open to video calls", "location", "openness to ideas" and "timezone". Fairly neat.
I was wondering why so complicated and how the model was learned...
> https://github.com/DevClad-Inc/devclad/blob/main/apps/server...
Aha, better indeed to rely on your own human learning ;-)
Is this like Lunchclub? https://lunchclub.com/
Similar. I took that as a starting point for the developer niche and I plan on turning it into something where developers can work on ideas together. I built an even more rough version of that here - https://connectdome.com (but then I realized I was dealing with feature-bloat) so now I'm experimenting a little slower.
OT, but that's an expertsexchange kind of domain...
> Every week, you are matched with another dev using an ML algorithm for a 1 on 1 call. I must admit I rolled my eyes when I read "ML algorithm" since it's usually marketing bullshit. Seemingly it's unsupervised knn w/ Scikit Learn, mixing variables like "open to video calls", "location", "openness to ideas" and "timezone". Fairly neat.
It indeed is. I'll try to find someone better at ML than me over time but I think this approach should serve well at least for the first 1k users or so if I'm not wrong.
This application will be able to read and write all user data. This includes the following:
Private email addresses
Private profile information
Followers
Earlier quoted context omitted.
It indeed is. I'll try to find someone better at ML than me over time but I think this approach should serve well at least for the first 1k users or so if I'm not wrong.
maybe just say “similarity” rather than “ml algorithm” since that is what you are doing