7 months ago
Beijing, ChinaMid Level / Senior
Responsibilities
- Build and scale AI inference infrastructure, including inference serving, scheduling, orchestration, and autoscaling.
- Design and optimize CPU/GPU resource management systems for utilization and cost efficiency.
- Implement production GPU scheduling and multiplexing technologies, including MIG, MPS, and virtualization.
- Optimize inference throughput, latency, and stability for high-concurrency and complex business scenarios.
- Contribute to system reliability, disaster recovery, observability, and cloud cost governance.
- Explore AI-native infrastructure and automated operations and maintenance systems.
Requirements
- 3–5 years of experience in backend or infrastructure engineering, with cloud-native or AI platform experience preferred.
- Strong proficiency in Go or Python and solid software engineering fundamentals.
- Deep understanding of Linux internals, networking, and distributed systems.
- Hands-on experience with Kubernetes, Docker, and microservices architecture.
- Prior project experience in inference systems, task scheduling, or resource management.
- Preferred experience includes GPU platforms, customized Kubernetes schedulers, Ray, model-serving frameworks, distributed inference, MIG, MPS, vGPU, SRE, observability, or cloud cost optimization.
- Open-source contributions or significant technical influence in the engineering community are advantageous.
Benefits
- Competitive salary, equity, and benefits package.
- Flexible work environment with remote and on-site options.
- Unlimited, flexible time off.
- Stock options available for core team members.
- 401(k) plan.
- Comprehensive health, dental, and vision insurance.
- Latest office equipment.
- Opportunities for professional growth and development.
Tech Stack
Categories
About Meshy LLC
Meshy LLC builds generative AI tools that convert text or reference images into usable 3D models for artists, game developers, and other creators. The company offers a web-based service and APIs for creating and editing 3D assets used in games, AR/VR, and visualization workflows. Headquartered in Sunnyvale, California, it is privately held and venture-backed by firms including Sequoia Capital and GGV Capital.