之

之江实验室 Zhejiang Lab Tech Stack

AI research lab focused on intelligent computing and materials science

Research Services Hangzhou, Zhejiang 1,001–5,000 employees Founded 2017 Nonprofit

Zhejiang Lab is a state-backed research institution founded in 2017 that applies AI — particularly foundation models and multimodal inference — to materials science and R&D workflows. The tech stack (PyTorch, Megatron-LM, DeepSpeed, Kubernetes, Hugging Face) reflects a research-heavy operation oriented toward large-scale model training and deployment. Pain points center on compute optimization, intelligent scheduling, and adapting general-purpose models to domain-specific scientific data, which maps directly to their active work on materials-focused inference platforms and elastic infrastructure.

Tech Stack 12 technologies

Core StackPyTorch Linux GitHub Hugging Face Kubernetes Jenkins GitLab CI/CD Grafana GPT Claude Megatron-LM DeepSpeed

What 之江实验室 Zhejiang Lab Is Building

◆Challenges

Optimizing computing resource allocation
Intelligent scheduling
Large-scale model training
Scaling petabyte-scale data
High-performance data retrieval
Implementing large models in vertical science scenarios
Adapting large models to scientific data
Large-scale system performance bottlenecks
High concurrency scheduling conflicts
Cross-region network communication failures

▲Active Projects

Deploy lightweight models into materials r&d
Materials multimodal inference model development
Platform scheduling r&d
Foundation model research and deployment
Ai-material science integration model
Unifying metadata and distributed storage architecture
Implementing elastic computing layer
Supporting continuous scaling and high-performance access
Multimodal reasoning models for materials
Ai inference framework

Hiring Activity

Accelerating15 roles · 15 in 30d

Department

Engineering

11

Research

4

Ops

2

Seniority

Senior

15

Junior

1

Mid

1

Company intelligence

Find more companies like 之江实验室 Zhejiang Lab by tech stack, pain points and active projects

Get started free

About 之江实验室 Zhejiang Lab

Zhejiang Lab operates as a nonprofit research institute in Hangzhou, China, with 1,001–5,000 employees. Its mission centers on intelligent computing as a strategic technology platform supporting China's innovation-driven development goals. The organization runs three core operational areas: applied AI research (foundation models and multimodal reasoning), platform engineering (scheduling, distributed storage, inference frameworks), and materials science integration. Current hiring is concentrated in engineering and research roles, with accelerating velocity and a senior-heavy workforce, indicating expansion of in-house model development and infrastructure capacity.

HeadquartersHangzhou, Zhejiang

Company Size1,001–5,000 employees

Founded2017

Hiring MarketsChina, United States

Frequently Asked Questions

What AI and ML tools does Zhejiang Lab use?

PyTorch, Megatron-LM, DeepSpeed, Hugging Face, GPT, Claude, and Kubernetes. Stack reflects focus on large-scale model training, inference optimization, and distributed deployment.

What is Zhejiang Lab working on?

Materials science AI integration (multimodal inference models, lightweight model deployment for materials R&D), foundation model research and deployment, and platform infrastructure (scheduling, distributed storage, elastic computing layers for high-performance access).