Latest AI JobsBrowse Jobs
4

LLM Engineer (Optimization)

42dotPangyo (Software Dream Center), South KoreaENGINEERING
Apply on company site
Salary
Not listed
Job type
Full-time
Work mode
Remote
Experience
Senior · 3+ yrs
Posted
22 Sept 2026
Status
Open · checked 7d ago

About this role

LLM Engineer (Optimization) is an open position at 42dot located in Pangyo, South Korea, focused on maximizing large language model inference performance across servers and edge environments. This role suits professionals with a background in machine learning infrastructure who want to design high-performance, low-latency, and cost-effective artificial intelligence systems.

Role Overview

LLM Engineer (Optimization) at 42dot centers on advancing large language model inference performance to meet real-world service requirements. The work involves improving model execution speed, reducing memory overhead, and researching efficient deployment strategies across multiple hardware platforms ranging from graphics processing units to edge devices.

Key Responsibilities

Enhance inference metrics such as latency, throughput, and memory efficiency for large language models within practical service deployments.
Build and refine inference engines and runtimes utilizing frameworks like vLLM, TensorRT-LLM, SGLang, llama.cpp, ONNX Runtime, and MLX.
Research model compression techniques including quantization methods like MXFP8, NVFP4, AWQ, and GPTQ, alongside pruning and distillation.
Design optimization tactics for mobile, embedded, and edge hardware architectures running CPU, GPU, or NPU hardware.
Perform profiling and benchmark evaluations using tools like NVIDIA Nsight Systems, Nsight Compute, and py-spy to resolve performance bottlenecks.

Requirements

Three or more years of professional background in large language models, machine learning infrastructure, or inference optimization.
Practical familiarity developing artificial intelligence inference engines or runtimes.
Understanding of graphics processing unit architecture, CUDA programming, or parallel computing principles.
Comprehension of model compression, quantization, and compiler optimization methodologies.
Working capability with deep learning frameworks such as PyTorch, ONNX, and TensorRT.
Proficiency in Python alongside C or C++ programming languages combined with software engineering capabilities.

Nice to Have

Hands-on deployment or engineering background with serving frameworks such as vLLM, TensorRT-LLM, SGLang, llama.cpp, MLX, or ONNX Runtime.
Experience authoring custom CUDA kernels, Triton kernels, or custom operators.
Familiarity with speculative decoding, prefill-decode disaggregation, KV cache compression, or expert parallelism.
Optimization background targeting NVIDIA, AMD, or edge processors including Orin, Thor, Qualcomm, and Apple Silicon.
Kubernetes container orchestration management for artificial intelligence serving or large GPU cluster maintenance.
Contributions to open-source software libraries within machine learning systems or peer-reviewed publications at academic venues like NeurIPS, ICML, ICLR, MLSys, OSDI, NSDI, or ASPLOS.

Skills and Tools

Python
C/C++
PyTorch
ONNX
TensorRT
vLLM
SGLang
llama.cpp
MLX
CUDA
Triton
TVM
MLIR
NVIDIA Nsight Systems
NVIDIA Nsight Compute
py-spy

About the Company

42dot operates out of the Software Dream Center in Pangyo, South Korea. The organization evaluates candidates through document screening, coding tests, a virtual initial interview, and a comprehensive second interview.

How to Apply

Interested candidates should submit their application directly on the official company careers page.

Skills and tech stack

PythonCC++PyTorchONNXTensorRTvLLMTensorRT-LLMSGLangllama.cppONNX RuntimeMLXCUDATritonTVMMLIRNVIDIA Nsight SystemsNVIDIA Nsight Computepy-spyLLMKubernetesMachine LearningFine-tuning

Questions candidates ask

Where is the LLM Engineer (Optimization) position located?

The role is located at the Software Dream Center in Pangyo, South Korea.

How many years of experience are required for this role?

Candidates must have at least three years of professional experience related to large language models, machine learning infrastructure, or inference optimization.

What programming languages are needed?

Applicants must be proficient in Python as well as C or C++.

What does the interview process consist of?

The evaluation process includes a document screening, a coding and assignment test, a remote first-round interview, and an in-person or virtual second-round interview.

Are there any specific exclusions requested for application materials?

Applicants should exclude resident registration numbers, family relations, marital status, salary history, photographs, physical attributes, and regional origins.

Ready to apply?

The application is on the official 42dot careers page.

Apply on company site