Algorithm AI System Researcher
FuriosaAI is a South Korean AI semiconductor company building energy-efficient inference accelerators, servers, and software for enterprise and cloud AI deployments.
Funding history
Investors
About FuriosaAI
FuriosaAI develops the RNGD AI inference accelerator and NXT RNGD Server, alongside a software toolchain for compiling, optimizing, and deploying LLM and agentic-AI workloads.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will identify and define research problems for next-generation NPU serving systems. You will use analytical modeling, roofline analysis, cost models, and simulators to evaluate architecture and algorithm design spaces. You will develop and validate serving algorithms through modeling, experimentation, and NPU implementations, advancing promising ideas into prototypes.
Requirements
- Master's degree or expected master's degree in AI/ML, computer architecture, or a related field, or equivalent experience
- Understanding of LLM inference and performance characteristics, including attention, KV cache, prefill, decode, batching, and multi-chip parallelism
- Ability to perform performance modeling, simulation, and data-driven analysis using Python and PyTorch
- Ability to structure ambiguous problems and lead hypothesis-validation cycles
- Ability to communicate analysis and technical decisions through clear documentation and presentations
Responsibilities
- Identify and define research challenges for next-generation NPU serving systems
- Explore serving architecture and algorithm design spaces using analytical modeling, roofline analysis, cost models, and simulators
- Develop and validate serving algorithms through modeling, experiments, and NPU-based implementations
- Advance promising ideas into proof-of-concepts and prototypes for productization