Here is how an engineer moonlighted as a matchmaker and pulled off a 100% match rate.
Working on Keeper, I dove headfirst into the wild world of matchmaking, confident that I could teach our proprietary LLM (Learning Love Model? Love Language Model? There’s something here ??) to tease out what causes the spark between two people. While I was confident machines could understand latent factors that drive long term relationship with the right data, I wanted to see how I would fare the old school way.
"Hey, how tough can it be? Match people with similar likes, and voilà!" Well, turns out, it’s not that simple. Shared interests in spaghetti or space operas don’t always equate to true love. Who would've guessed? (Spoiler: Almost everyone but me!)
After a few comical missteps, my inner geek pushed me to hit the books. I wasn’t just reading cheesy love stories; I was diving into the nitty-gritty of psychological studies and ancient matchmaking wisdom (yes, that’s a thing). My mission? To crack the code of long term relationships. Fast forward through a montage of research, trial, and error, and there I was, blending personality insights, emotional IQ, and even a pinch of astrological magic (because why not?) to find that perfect match. I actually did it – I found a match that seemed as perfect as a scripted rom-com finale. I was ready to take a bow, thinking I’d nailed the perfect love algorithm.
With a mix of pride and excitement, I presented this perfect match according to me to our senior matchmaker, expecting cheers and confetti. I was sure that my recent research had allowed me to pair two perfect soulmates. But what actually happened when the matchmaker pulled up the match on our portal (with a knowing smirk, I might add), my match not did not pop up as the the ultimate match (again, to my surprise but no one else’s). Our senior matchmaker in conjunction with our AI not only found the perfect match that I would never have considered on my own but also flagged all the potential disasters I had blissfully ignored in my proposed match!
Armed with this newfound wisdom, I presented the match suggested by our senior matchmaker + AI as my own ?? . And it was a home run!
So, now comes the key part - quit on that high note ??.
Cue the credits: I swapped my matchmaker's hat for my trusty engineer’s keyboard, but not without some key takeaways:
???? The criteria for finding a match in the digital age are among the most long tail searches in any industry, revealing the intricate nature of human relationships.
???? Human experts have been able to find the “secret sauce” to matchmaking, but in a two sided network like matchmaking, a human matchmaker would be the bottleneck in truly achieving any network effects.
???? A harmonious blend of expert human judgment and advanced AI proves to be the most effective approach, outperforming either in isolation.
So if you're looking for a keeper, you might want to check out Keeper.
