O
OkCupid Tech Blog
tech.okcupid.com · Company blog · on HN since 2015
Engineers at OkCupid write about search, matching, A/B testing, image hashing, GraphQL, and Rust.
What HN saysHN commenters cite it for its Vespa versus Elasticsearch comparison and its work on perceptual image hashes.
26HN people shared it
4linked it in comments
42stories submitted
631story points
4 years agolast shared
2015people per year2026
Most-shared posts
- 1OkCupid Presents: Vespa vs. Elasticsearch for matching millions of people5 people · 19 points · first shared 2020 · discussion on HN
- 2Swiping Right on Bloom Filters4 people · 14 points · first shared 2015 · discussion on HN
- 3The pitfalls of A/B testing in social networks3 people · 132 points · first shared 2017 · 18 comments on HN
- 4Evaluating Perceptual Image Hashes at OkCupid3 people · 105 points · first shared 2017 · 14 comments on HN
- 5Moving OkCupid from REST to GraphQL3 people · 44 points · first shared 2020 · 21 comments on HN
- 6When Debuggers Lie by OkCupid2 people · 10 points · first shared 2017 · discussion on HN
- 7Why we decided against GraphQL for local state management1 people · 124 points · first shared 2020 · 95 comments on HN
- 8Rust and RPC at OkCupid1 people · 63 points · first shared 2017 · 10 comments on HN
- 9A failed dating experiment – Responsiveness Sort1 people · 39 points · first shared 2018 · 9 comments on HN
- 10How OkCupid predicts who you will like1 people · 34 points · first shared 2017 · 7 comments on HN
- 11How OkCupid organizes its multi-page React app1 people · 11 points · first shared 2017 · discussion on HN
- 12Building an Android Thin-Client at OkCupid1 people · 6 points · first shared 2018 · discussion on HN
- 13An Intro Guide to Protocol vs. Object Oriented Programming1 people · 3 points · first shared 2020 · discussion on HN
- 14Getting Hooked on React Hooks1 people · 3 points · first shared 2019 · discussion on HN
- 15The Shadow Request: Troubleshooting OkCupid’s First GraphQL Release1 people · 3 points · first shared 2020 · 2 comments on HN
What HN people wrote when they linked it
vitno · 2017-04-21
we also have a tech blog that we are revitalizing!
http://tech.okcupid.com
363849473754 · 2022-11-20
Has anyone ever had success on an app like Hinge, then deleted it only to see a large drop in your success rate when nothing else changed about your profile?
If so, it could be due to things like this:
https://tech.okcupid.com/evaluating-perceptual-image-hashes-at-okcupid…
Match Group owns OkCupid, Hinge, Tinder and so on.. so wouldn’t be surprising if they all used the same or similar algorithms for spam prevention and noob boost abuse.
They probably also collect other data like your phone number, deviceID, etc.
This is bad because if you ever delete your profile and remake it then your internal ranking could plummet due to these aggressive spam filters.
zmjjmz · 2022-02-23
We've had good success with Vespa [1] for this usecase, both using it to rank by vector similarity and apply business logic oriented filters.
Search latencies are fairly low, even with constantly updating documents.
[1] https://tech.okcupid.com/vespa-vs-elasticsearch-for-matching-millions-…
pippy360 · 2017-08-10
I think for handling truly/non-affine arbitrary transformations we will have to resort to ML. Then we could have matching very similar to how humans do it (where we really don't care if the transformation is affine/non-affine we just care if it's a huge transformation the makes the image unrecognisable). But I really don't know much about ML.
Okcupid talked about this in an article about image hashing [0] and they have a nice quote:
"The end-to-end approach of training a Convolutional Net to embed images such that the positive / negative case distances separate nicely is probably the cool new hypebeast way forward. This would have the added advantage of being able to handle whatever transformations you can think of at training time."
[0]https://tech.okcupid.com/evaluating-perceptual-image-hashes-okcupid/