Skip to work
← All work

KLUTCH

Gig-economy matching run on local models, so the matching logic does not depend on somebody else's API bill.

2nd runner-up of 30+ teams at DLSU Hackercup, August 2025

August 2025 · 2nd runner-up, 30+ teams

runner-up
2ndrunner-upout of 30+ teams at DLSU Hackercup

The idea

Gig work matching is a ranking problem wearing a marketplace costume. KLUTCH approached it as an AI-integrated matching platform for gig-economy work, and I architected it and pitched it to a panel of industry experts.

The mechanism

Matching ran on local models through Ollama rather than a hosted API. Under hackathon conditions that is a practical decision as much as a technical one (no key, no quota, no latency to a third party), and it makes the cost structure of the thing you are pitching honest.

The result

Second runner-up out of more than 30 teams at the De La Salle University Hackercup, August 2025. A competition build, judged on architecture and pitch.

Screenshot

KLUTCH's search home: trade categories from electrician to carpenter, saved workers with hourly rates in pesos, and a recently viewed profile with its rating and city.
KLUTCH's search home: trade categories from electrician to carpenter, saved workers with hourly rates in pesos, and a recently viewed profile with its rating and city.

More work