BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260417T190002Z
LOCATION:West Building\, Exhibit Hall C
DTSTART;TZID=America/Los_Angeles:20250810T080000
DTEND;TZID=America/Los_Angeles:20250810T090000
UID:siggraph_SIGGRAPH 2025_sess618_job_130@linklings.com
SUMMARY:Research Scientist-LLM System
DESCRIPTION:We are looking for a Research Scientist to design and optimize
  Huawei’s large model infrastructure, enabling highly efficient training, 
 inference acceleration, and scalable deployment. This role focuses on buil
 ding cutting-edge system architectures and engineering platforms that empo
 wer large model research and production at scale. You will work at the int
 ersection of AI infrastructure, distributed systems, and high-performance 
 computing, driving innovations that accelerate Huawei’s foundation model e
 cosystem.\n\nResponsibilities:\n1.Lead system-level optimizations for larg
 e model training, including but not limited to parallel strategies (PP/TP/
 CP), memory optimization, communication optimization, and fault tolerance 
 in large-scale training clusters.\n2.Drive inference acceleration for larg
 e models, including but not limited to fused operator optimization, parall
 el strategy optimization, memory optimization, communication optimization,
  and other parallel or mathematical optimization techniques.\n3.Design, de
 velop, and maintain large model training platforms and toolchains to provi
 de simple, efficient, and scalable training systems for algorithm teams. T
 his includes but is not limited to automated evaluation systems, job monit
 oring systems, and visualization tools.\n4.Build and maintain large model 
 inference and deployment platforms, including automated pipelines, one-cli
 ck deployment systems, and data feedback systems, to enable rapid and cost
 -effective internal and external deployment of large models.\n\n
END:VEVENT
END:VCALENDAR