Software Engineer Machine Learning Engineer Agentic AI
FuriosaAIVisit FuriosaAI website
FuriosaAI is a South Korean AI semiconductor company building energy-efficient inference accelerators, servers, and software for enterprise and cloud AI deployments.
Seoul, South Korea
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 design and implement autonomous agent systems that explore, execute, evaluate, and improve engineering solutions. You will develop multi-agent workflows, automated evaluation systems, post-training experiments, and production-quality integrations for engineering workflows.
Requirements
- Bachelor's degree in Computer Science or a related field
- Model training or experimentation using PyTorch, JAX, TensorFlow, or another ML framework
- Experience implementing or experimenting with LLM or Agentic AI systems
- Ability to understand research papers and turn research into implementations
- Agent system development
- Tool use
- Reasoning
- Memory
- Planning
- Multi-agent orchestration
- SFT
- Reinforcement learning
- Preference optimization
- Agent evaluation
- Benchmark design
- LLM-as-a-judge
- Automated evaluation
- LangGraph
- AutoGen
- Google ADK
- Distributed training
- Inference optimization
- Large-scale experimentation
Responsibilities
- Research and implement agent planning, tool use, memory, reasoning, and self-improvement mechanisms
- Develop multi-agent workflows and orchestration
- Build systems that execute, validate, evaluate, and iteratively improve agent outputs
- Evaluate and optimize agents using LLM-as-a-judge, task-specific metrics, and benchmarks
- Experiment with and apply SFT and reinforcement learning post-training techniques
- Apply Agentic AI and post-training research to engineering problems