With the fast growth of ByteDance's business, ByteDance's system infrastructure is currently at a massive scale and requires versatile system solutions.
Our team works on advanced R&D projects, focusing on
LLM/AI + Infrastructure technologies, which includes both infrastructure for
LLM/AI and
LLM/AI for infrastructure.
To name a few, our TextToSQL project ranks top on well known industry benchmarks.
We also work on advanced multimodal data processing technologies using LLM, such as scalable and efficient semantic operators and AI functions.
Besides achieving great business impacts, we also encourage publishing on top tier conferences.
In year 2025 alone, our team published nearly 20 papers in top tier conferences, such as SIGMOD, VLDB, FSE, ICLR, EuroSys, WWW etc. We hire students with great technical skills, willingness to learn and solve complex technical challenges and passion in making an impact on millions of users.
With the large-scale deployment of large language models (LLMs) and AI Agents, traditional cloud-native infrastructure can no longer meet the extreme performance and elasticity demands of AI workloads.
This project conducts systematic research across the full stack of AI infrastructure, focusing on the following areas: Data Management for LLM and Agents
1. Multi-modal query processing: Support seamlessly integrated multi-modal query processing, including vector, full-text and regular SQL query processing over various typed data.
In addition, how to support large scale semantic operators in a cost-effective and low-latency way is also our research focus.
Intelligence & Agent Architecture
- Explore infrastructure auto-optimization based on AI Agent workflows.
Build a self-evolving business Agent framework, enabling full-stack intelligent optimization through “AI for Infrastructure”.
We look into various ways to apply
AI/LLM in solving infrastructure problems, such as NL-to-SQL, Auto Skills etc. This project aims to build next-generation AI-native infrastructure to support LLMs and AI Agents, improving resource utilization, reducing costs, enabling elastic scalability, and driving the evolution of AI infrastructure technologies.
Topic Content:With the large-scale adoption of LLMs and AI agents, traditional cloud-native infrastructure can no longer meet the ultra-high performance and elasticity requirements of AI workloads.
This topic conducts systematic research across the entire AI infrastructure stack:
1. Network and Observability: Research intelligent fault localization and root cause analysis for large-scale AI clusters, combined with intelligent tuning of time-series databases to improve cluster stability.
2. Storage Systems: Develop serverless high-performance elastic file systems and storage acceleration architectures specifically for AI scenarios, explore hardware-software co-optimization for DPU, and overcome AI storage performance bottlenecks.
3. Data Center Power Scheduling: Research
GPU/CPU/MEM heterogeneous collaborative scheduling technologies, build a heterogeneous power orchestration system for AI agents, and address scheduling challenges including heterogenous workloads and state dependencies.
4. Vector Retrieval: Optimize core vector retrieval technologies for LLM-powered applications, building a cloud-native distributed vector index engine to meet ultra-large-scale vector retrieval demands with low latency and low cost.
5. Intelligence and Agent Architecture: Explore automatic infrastructure optimization based on AI Agent workflows, build a self-evolvable business agent framework, and enable full-stack intelligent optimization through AI for Infra.
This topic aims to build a next-generation AI-native infrastructure to support the deployment of LLMs and AI agents, improve resource utilization, reduce costs, support elastic scaling, and drive the technological evolution of AI infrastructure.