Teammate Finder
Context
Leetify is a performance analytics platform for CS2 players. The core product helps players understand their mistakes, track stats, and improve. What it didn’t have was a way to find people to play with — which turns out to be one of the top reasons players quit.
I noticed this gap through support tickets, Discord discussions, and my own time playing on the platform. No one asked me to solve it. I decided to.
The problem
Players who want to improve need consistent practice partners who are at a similar skill level and take the game seriously. Finding those people through random matchmaking is unreliable. The existing workaround — posting in subreddits and Discord servers — is fragmented, slow, and outside the product entirely.
The question I started with: If Leetify already knows everything about how I play, why can’t it just find me a teammate?
What I designed
A matching flow built on top of Leetify’s existing data. Players set a short availability window and a few preferences — playstyle, rank range, mic preference. The system surfaces compatible players and lets both sides accept a match.
The key differentiator: once two players matched, a Steam bot automatically created the lobby and sent both a join link. No copy-pasting friend codes. No waiting to see if the other person actually shows up. Zero manual steps.
Process
Started with Discord. Spent two weeks reading player complaints and asking questions in the Leetify community server. Ran five informal interviews — screenshare, casual, 20 minutes each. The pattern was consistent: people wanted curation, not a search box.
From there I mapped the existing data model. Leetify already tracked rank, playstyle tendencies, communication patterns, and activity windows. The matching logic was essentially already there — it just wasn’t surfaced as a product.
Ran three rounds of concept testing with community members. The Steam bot concept got the strongest reaction — “wait, it actually creates the lobby for you?” was a consistent response. That told me the automation was the real value, not the matching itself.
Why it wasn’t shipped
Roadmap prioritization, not product quality. The feature was fully designed and handed off, but the team shifted focus to a major data pipeline rebuild. It lives in Figma, ready to go.
What I learned
Proposing work without a brief is a different skill than executing a scoped request. You have to be your own product manager, your own researcher, and your own advocate. The hardest part isn’t designing the feature — it’s convincing the room it’s worth the engineering time.