Algorithm - AI System Researcher
New roles land in Following.
Seoul, South KoreaPosted 1 month ago
What you'd do
- Identify and define major research challenges in next-generation cluster-level NPU serving systems.
- Explore serving architectures using analytical modeling,roofline analysis,cost models,and internal simulators.
- Model and validate scheduling,batching,KV cache,and disaggregation algorithms via experiments and NPU implementation.
What they want
- Hold or expect a master's degree in AI,ML,computer architecture,or equivalent experience.
- Understand LLM inference mechanics including attention,KV cache,prefill/decode,batching,and multi-chip parallelism.
- Perform performance modeling,simulation,and data analysis using Python and PyTorch proficiently.
Nice to have
- Bring experience analyzing or optimizing distributed systems,cluster software,or large-scale workloads.
- Understand internals of LLM serving frameworks such as vLLM,SGLang,or TensorRT-LLM.
- Apply roofline analysis,analytical performance models,or total cost of ownership evaluation experience.
Seoulstart's read
- Visa tier (likely)
- E-7 specialty / professional (likely)
- Korean language
- Korean requirement not stated