Hao Zhang

Research Scientist at ByteDance

Context infrastructure for agents.

I am a Research Scientist at ByteDance. I expect intelligence to become increasingly abundant, with much of it instantiated as long-lived agents that collaborate with each other and with humans. As these agents continuously read from and write to shared information environments, organizing that information—and turning it into the right context for each task—will become a fundamental systems problem. I build self-improving context infrastructure toward that future.

My current work focuses on the systems between agents and external information: how information is integrated, organized, indexed, retrieved, and assembled into task-specific context.

AutoIA @ ByteDance is the platform for this work, using task-level feedback to optimize retrieval pipelines first and, when needed, the underlying information environments they rely on.

Longer-term view

Publications contains the full research record.

  • 30+ publications in database systems and AI, including SIGMOD, VLDB, ICDE, EMNLP, COLM, TKDE, and The VLDB Journal.
  • #1 on both tracks of LDBC SNB Interactive, the leading audited benchmark for graph database systems: Declarative, 2024 (3,000× over #2) and Imperative, 2025.

I received my Ph.D. from the Chinese University of Hong Kong, advised by Prof. Jeffrey Xu Yu and Prof. Hong Cheng, and my B.S. in Computer Science from Wuhan University.

  • ByteDance Soaring Star Talent Program
  • Huawei Gold Award for Individual Excellence — Top 1%
  • Huawei TopMinds Talent Program — First Class (highest tier)

I welcome collaboration on context infrastructure, retrieval systems, and graph/vector systems. I am also recruiting research interns at ByteDance in Shenzhen; email zhanghaowuda12 [at] gmail [dot] com with [Intern] in the subject.

  • 08/2026DocNavRAG released on arXiv.
  • 07/2026AdaMM released on arXiv.
  • 07/2026Sema accepted by PVLDB’26; CoreSemDB accepted by COLM’26.

→ All news