Latest AI JobsBrowse Jobs
vessl-ai logo

AI Research Engineer (Senior)

vessl-aiSeoulEngineering
Apply on company site
Salary
Not listed
Job type
Full-time
Work mode
Remote
Experience
Senior · 4+ yrs
Posted
29 Sept 2026
Status
Open · checked 7d ago

About this role

This senior research role suits an experienced machine learning professional based in Seoul who wants to lead advanced optimization and post-training initiatives for large language models. The position is located in Seoul and involves shaping the research roadmap for GPU infrastructure and AI agent workloads at VESSL AI.

Role Overview

This position sits within the research and development division of VESSL AI, a GPU cloud organization operating in Seoul. The team focuses on agentic artificial intelligence, reinforcement learning post-training, and distributed cache or accelerator kernel optimization across modern graphics processing units ranging from H200 to B300 and GB300. The senior engineer acts as a core team member by defining problems, executing frontier research, and feeding performance improvements directly into products and customer experiences.

Key Responsibilities

Investigate and improve inference efficiency on large GPU clusters using frameworks like vLLM and SGLang, focusing on multi-token prediction, speculative decoding, lossless compression, and kernel optimization.
Design multi-node post-training pipelines utilizing open frameworks, implementing reinforcement learning and distillation across reasoning, math, and coding domains.
Translate benchmarking and model validation findings directly into production inference and post-training features.
Drive research roadmaps, lead code and architectural reviews, document technical workflows, and mentor junior engineers.
Review recent papers and open-source releases to share insights through internal seminars and studies.

Requirements

Hold a doctoral degree in computer science, electrical engineering, AI, machine learning, or a related field, or possess a master's degree paired with at least four years of practical or research experience.
Demonstrate measurable inference performance gains via speculative decoding, lossless compression, or kernel optimization such as CUDA or Triton.
Exhibit proven experience improving model performance through fine-tuning or post-training of large language models.
Read recent papers quickly and convert core ideas into implemented, validated code.
Direct technical decision-making and research roadmaps within team projects or laboratories.
Communicate fluently in English at a business level.

Nice to Have

Authored publications in international machine learning or systems conferences like NeurIPS, ICML, or MLSys.
Contributed to open-source projects including vLLM, SGLang, verl, or slime.
Understand reinforcement learning algorithms, flash attention, quantization techniques, and continuous batching.
Experience recruiting, onboarding, and mentoring researchers or engineers.
Familiarity with operating multi-node GPU clusters via Slurm or Kubernetes.
Willingness to cross boundaries between software engineering, infrastructure, and products when solving complex issues.

Skills and Tools

Python
CUDA
Triton
vLLM
SGLang
verl
slime
Kubernetes
Slurm

Salary and Benefits

Annual self-development support budget of up to 1.2 million KRW.
Annual global market exposure opportunity through a US experience pass.
Support for physical book purchases and electronic library access.
Flexible working hours with a core start time choice between 8:00 AM and 11:00 AM.
Monthly all-hands meetings, team gatherings, overtime meal, and taxi support.
Personal snack budget, long-term service refresh leave, and annual comprehensive health checkups.
Birthday half-day leave, congratulatory gifts, holiday presents, family event support, and childbirth leave funds.

About the Company

VESSL AI provides GPU infrastructure and cloud services to enterprise AI clients through GPU-as-a-Service, assisting teams in securing stable computing resources. Serving domestic and international enterprises, startups, and academic labs across the United States, the organization scales rapidly within the global AI infrastructure market.

How to Apply

Apply on the official company careers page by submitting a detailed career history resume and a portfolio or Git repository link.

Skills and tech stack

PythonCUDATritonvLLMSGLangverlslimeKubernetesSlurmLLMReinforcement LearningFine-tuning

Questions candidates ask

Where is the job located?

The position is based in Seoul, South Korea.

What qualifications are required for education and experience?

Applicants must hold a PhD in a relevant technical field, or a Master's degree with at least 4 years of practical or research experience.

What language proficiency is expected?

Candidates must have business-level English communication skills.

What benefits are provided for personal growth?

The company offers up to 1.2 million KRW annually for self-development, book purchase support, and an annual US market experience pass.

What is the interview process like?

The hiring pipeline includes a document screening, take-home assignment, technical interview, resume and culture interview, and a CEO interview, followed by a reference check for experienced hires.

Ready to apply?

The application is on the official vessl-ai careers page.

Apply on company site