V

VESSL AI

GPU cloud platform for AI workloads

Senior AI Research Engineer

AI ResearchConfirmed live in the last 24 hours
Full-timeSeoul, South KoreaNo salary listedHybrid

About the Job

Requirements

  • Hold a PhD in computer science, electrical and electronic engineering, AI/ML, or a related field, or hold a master's degree and have at least four years of practical or research experience.
  • Have demonstrated experience achieving meaningful inference performance improvements in at least one of speculative decoding, lossless compression, or kernel optimization using CUDA, Triton, or similar technologies.
  • Have experience improving model performance through fine-tuning or post-training of large language models.
  • Be able to quickly read recent research papers and implement and validate their core ideas in code.
  • Have experience leading research roadmaps or technical decisions in a team project or research lab.
  • Have business-level English communication skills.

Responsibilities

  • Improve inference efficiency of framework-based engines such as vLLM and SGLang in large-scale GPU cluster environments, focusing on multi-token prediction, speculative decoding, lossless compression, and kernel optimization using CUDA, Triton, or similar technologies.
  • Design multi-node large-scale post-training pipelines using verl, vime, and slime, and implement and optimize reinforcement learning and on-policy/off-policy distillation across reasoning, math, and code domains.
  • Translate research outcomes into VESSL inference and post-training product features and real customer workloads through benchmarking and model validation.
  • Lead research-priority and roadmap discussions; strengthen the team's research capabilities through code and design reviews, technical documentation, and mentoring junior engineers.
  • Continuously review current papers and open-source projects in agentic AI and LLM serving and training, and share insights through internal study sessions and seminars.

Desired Qualifications

  • Have published papers at international machine-learning or systems conferences such as NeurIPS, ICML, or MLSys.
  • Have contributed to open-source projects such as vLLM, SGLang, verl, or slime.
  • Have a deep understanding of reinforcement-learning algorithms such as GRPO and PPO, and current inference optimization techniques such as FlashAttention, quantization formats MXFP4, NVFP4, and W4AFP8, and continuous batching.
  • Have contributed to team growth through researcher or engineer hiring, onboarding, or mentoring.
  • Understand operation of multi-node GPU clusters, including Slurm or Kubernetes.
  • Be willing to work hands-on across software engineering, infrastructure, and product areas when needed to solve problems beyond a defined scope.
  • Enjoy independently investigating ambiguous problems with no predetermined answer until a solution is found.

About the Company

VESSL AI

GPU cloud platform for AI workloads

What VESSL AI does: It runs VESSL Cloud, a GPU cloud platform built for AI workloads. It gives teams on-demand access to high-end GPUs (A100, H100, B200, B300, GB200) to run activities like LLM fine-tuning, large-scale model training, and AI inference. How it works: Users launch GPU instances in VESSL Cloud from VESSL’s own data center infrastructure, paying per minute for usage, and using the platform to accelerate AI development without managing physical hardware. How it stands out: It operates its own data center and offers per-minute billing, delivering up to 80% cost savings versus AWS and GCP, focusing specifically on AI workloads rather than general-purpose cloud services. What the goal is: To help AI teams move quickly from idea to production by removing infrastructure overhead and providing scalable, affordable GPU resources for AI work.

Company size
51-200
Company stage
Series A
Total funding
$16.6M
Headquarters
San Francisco, CA
Founded
2020

Recently Posted Jobs

Explore more opportunities at VESSL AI

See more jobs