BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260202T201224Z
LOCATION:Hall 6
DTSTART;TZID=America/Chicago:20251117T165400
DTEND;TZID=America/Chicago:20251117T165400
UID:submissions.supercomputing.org_SC25_sess553_job293@linklings.com
SUMMARY:Senior Machine Learning Engineer, Post Training & Speculative Deco
 ding
DESCRIPTION:Mission:\nWe are seeking a highly skilled Machine Learning Eng
 ineer to join our advanced model development team. This role focuses on pr
 e-training, continued training, and post-training of models, with a partic
 ular emphasis on draft model optimization for speculative decoding and qua
 ntization-aware training (QAT). The ideal candidate has deep experience wi
 th training methodologies, open-weight models, and performance-tuning for 
 inference.\n\nResponsibilities & opportunities in this role:\nLead pre-tra
 ining and post-training efforts for draft models tailored to speculative d
 ecoding architectures.\nConduct continued training and post-training of op
 en-weight models for non-draft (standard) inference scenarios.\nImplement 
 and optimize quantization-aware training pipelines to enable low-precision
  inference with minimal accuracy loss.\nCollaborate with model architectur
 e, inference, and systems teams to evaluate model readiness across trainin
 g and deployment stages.\nDevelop tooling and evaluation metrics for train
 ing effectiveness, draft model fidelity, and speculative hit-rate optimiza
 tion.\nContribute to experimental designs for novel training regimes and s
 peculative decoding strategies.\n\nIdeal candidates have/are:\n5+ years of
  experience in machine learning, with a strong focus on model training.\nP
 roven experience with transformer-based architectures (e.g., LLaMA, Mistra
 l, Gemma).\nDeep understanding of speculative decoding and draft model usa
 ge.\nHands-on experience with quantization-aware training, including PyTor
 ch QAT workflows or similar frameworks.\nFamiliarity with open-weight foun
 dation models and continued/pre-training techniques.\nProficient in Python
  and ML frameworks such as PyTorch, JAX, or TensorFlow.\n\nPreferred Quali
 fications:\nExperience optimizing models for fast inference and sampling i
 n production environments.\nExposure to distributed training, low-level ke
 rnel optimizations, and inference-time system constraints.\nPublications o
 r contributions to open-source ML projects.\n\nAttributes of a Groqster:\n
 Humility - Egos are checked at the door\nCollaborative & Team Savvy - We m
 ake up the smartest person in the room, together\nGrowth & Giver Mindset -
  Learn it all versus know it all, we share knowledge generously\nCurious &
  Innovative - Take a creative approach to projects, problems, and design\n
 Passion, Grit, & Boldness - no limit thinking, fueling informed risk takin
 g\n\nRegistration Category: Technical Program Reg Pass, Workshop Reg Pass,
  Tutorial Reg Pass, Exhibits Reg Pass\n\nCompany: Groq\n\nIn-Person / Remo
 te: Remote\n\nPart Time / Full Time: Full Time\n\nPosition Type: Permanent
 \n\n
END:VEVENT
END:VCALENDAR