<ai/ml · weekly>

GPU Ghosts UTHM

small models, big inference. we run LLMs on a single 3090 in the FSKTM cluster.

lead · @iman.fkmp · 23/30 members · recruiting

<current mission>

open-source a sub-200ms Bahasa Melayu speech-to-text server as a UTHM capstone.

<code of conduct>

  • · ship something every sprint. anything. even a typo fix.
  • · review at least 1 PR that isn't yours.
  • · don't be a jerk. we have plenty of room for sharp opinions and zero room for rudeness.
  • · public channels, public work. unless it's a private security concern.

<next sessions>

next

Fri 5 Jun

13:00 MYT

+1 week

Fri 12 Jun

13:00 MYT

+2 weeks

Fri 19 Jun

13:00 MYT