Role overview
Own end-to-end delivery of AI agents and compute appliances for state-owned enterprise customers, coordinating presales, algorithms, infrastructure and operations teams through project acceptance.
Responsibilities
- Deploy LLM applications on edge hardware and domestic NPU devices at customer sites, including environment setup and version releases.
- Integrate hardware and software on site, adapt models and optimize inference, and troubleshoot hardware, software, network and compute issues.
- Conduct proof-of-concept tests and on-site demonstrations, and follow through on delivery and customer acceptance.
- Document deployments and failure cases, write scripts to improve delivery efficiency, and relay issues to internal engineering teams.
Requirements
- Two to four years of deployment and implementation experience in Linux environments; familiarity with Docker.
- Understand LLM inference deployment; experience with vLLM, Triton or domestic NPUs is preferred.
- Network troubleshooting skills, Shell scripting and basic Python debugging.
- Available for short domestic business trips and able to communicate effectively at customer sites.
Nice to have
- Experience with edge LLM optimization, domestic compute platforms or proof-of-concept delivery.