Skip to main content
Silan Hu

NUS · Computing / AI Systems

Silan Hu

NUS PhD · Emerging AI systems researcher for executable agent infrastructure

Turning agent memory, execution semantics, and knowledge systems into a research program for reliable executable AI agents.

silan.hu@comp.nus.edu.sgSingapore🇸🇬 / Beijing, China🇨🇳

About Me

I am an NUS PhD student advised by Prof. Xiaokui Xiao, building the systems layer that makes uncertain AI execution stable and governable. My work sits where databases, agent runtime infrastructure, and machine learning systems meet: how an agent stores procedural knowledge, retrieves the right context, executes with transactional discipline, and improves from previous attempts.

I came to this problem through both research and production systems: GEM-Bench, my first-author SIGKDD 2026 CCF-A paper on generative engine marketing benchmarks; VDSAgents, a journal paper on PCS-guided multi-agent data-science automation; FOKE, my first-author work on personalized explainable education; and production experience optimizing infrastructure for Open-Sora 2.0, a former open-source SOTA text-to-video model, plus backend development for a commercial text-to-video platform serving 300,000 users. I was admitted to the NUS PhD program with a full scholarship through the NUSGRTII innovation and entrepreneurship track.

My current research agenda is deliberately narrow and compounding: AI-native databases for agents, latent state and procedural memory, verifiable multi-agent workflows, and decentralized capability invocation. On the systems side, I am building EasyRemote / EasyNet, an open-source network for agent execution and ability calls across devices. The long-term goal is a credible substrate for executable AI agent networks — a layer agents can use to remember, coordinate, and act.

Recent Moments

May 2026

A PhD is driving one nail into the wall
Activelast month

I increasingly think doing a PhD is like driving one nail into a wall—the wall
being the accumulated wall of human knowledge.

At first, you may not even see the wall. You do not yet know what counts as a
nail, where one could be placed, how the existing nails relate to each other,
or which gaps are real. Reading is partly the work of learning to see that
surface.

Then comes another problem: choosing where your nail should go. The value of
a nail depends on the wall around it, but its shape also comes from your own
judgment. You need enough basic ability to see prior nails, understand how they
were driven, and recognize the rules of the game. Yet the reason for choosing
this particular nail cannot be borrowed entirely from other people.

So the PhD may be less about covering an area than about making one
well-founded intervention: seeing the wall, locating a gap, shaping a nail, and
being able to explain why it belongs there.

This understanding is probably only provisional—a stage-specific view.

silan-viking is moving along
Ongoing2 months ago

silan-viking — the personal context system behind this site — has been making good progress lately. The Rust engine and CLI are coming together, with content sync and publishing now working end to end.

Education

  1. National University of Singapore logo

    PhD in Computer Science

    2026-01 - Present
    • Full scholarship; admitted via the NUSGRTII innovation & entrepreneurship program.
    • Advised by Prof. Xiaokui Xiao (NUS).
    • Research agenda: AI-native databases for agents, procedural memory, execution semantics, and verifiable multi-agent orchestration.
    • Positioned at the intersection of database systems, AI agent infrastructure, and machine learning systems.
  2. National University of Singapore (NUS) logo

    Master of Computing (AI Specialisation)

    2024-08 - 2025-12
    • School of Computing — Artificial Intelligence specialisation.
    • Transitioned into the PhD program in January 2026 via the NUSGRTII program.
    • Built the bridge from applied AI systems into a research program on reliable agent infrastructure.
  3. Macau University of Science and Technology (MUST) logo

    Bachelor of Science in Computer Science (AI Minor)

    2020-09 - 2024-06
    • Received full scholarship of 250,000 CNY for Master's program.
    • Dean's List Student.
    • President of the Computer Science and Engineering Student Association (source).
    • Core Courses: Graduation Project (A+), Machine Learning (A+), Computer Programming (A+), Data Structures (A), Digital Logic (A+), Database Systems (A+), etc.

Publications

VDSAgents: A PCS-Guided Multi-Agent System for Veridical Data Science Automation

VDSAgents: A PCS-Guided Multi-Agent System for Veridical Data Science Automation

journalStat 15(1): e701262026-01

A five-agent pipeline (Define → Explore → Model → Evaluate, with a PCS guide) that automates an end-to-end data-science workflow without giving up reproducibility.

Authors: Yunxuan Jiang, Silan Hu, Xiaoning Wang, Yuanyuan Zhang, Xiangyu Chang

Open-Sora 2.0: Training a Commercial-Level Video Generation Model in $200 k

Open-Sora 2.0: Training a Commercial-Level Video Generation Model in $200 k

preprintarXiv preprint arXiv:2503.096422025-03

An open-source video diffusion model trained for $200k — proving you can hit commercial-grade quality on a startup budget by branching denoising trajectories and picking the best with a verifier.

Authors: Zangwei Zheng, Xiangyu Peng, Yuxuan Lou, Chenhui Shen, Tom Young, Xinying Guo, Binluo Wang, Hang Xu, Hongxin Liu, Mingyan Jiang, Wenjun Li, Yuhui Wang, Anbang Ye, Gang Ren, Qianran Ma, Wanying Liang, Xiang Lian, Xiwen Wu, Yuting Zhong, Zhuangyan Li, Chaoyu Gong, Guojun Lei, Leijun Cheng, Limin Zhang, Minghao Li, Ruijie Zhang, Silan Hu, Shijie Huang, Xiaokang Wang, Yuanheng Zhao, Yuqi Wang, Ziang Wei, Yang You

FOKE: A Personalized and Explainable Education Framework Integrating Foundation Models, Knowledge Graphs, and Prompt Engineering

FOKE: A Personalized and Explainable Education Framework Integrating Foundation Models, Knowledge Graphs, and Prompt Engineering

conferenceChina National Conference on Big Data and Social Computing2024-01

China National Conference on Big Data and Social Computing, pp. 399–411

A personalised education framework that fuses an LLM with a knowledge-forest graph of the learner, so explanations land at the right depth for the right student.

Authors: Silan Hu, Xiaoning Wang

S
2026Public portfolio

silan-viking — Personal Context System

The content engine behind this website — a schema-governed personal context system that treats a half-formed idea and a published essay as the same kind of object at different lifecycle stages.

content-systemmcppersonal-knowledgerust

Work Experience

  1. NUS Computing logo

    Research Assistant

    2025-08 - 2025-12
    NUS ComputingSingapore
    • Research Assistant, NUS School of Computing, supervised by Prof. Xiaokui Xiao.
    • Conducted research on GEM-Bench, accepted at KDD 2026.
    • Admitted to the Ph.D. program (January 2026) with a full scholarship through the NUS Graduate Research Talent & Innovation Initiative (NUSGRTII).
  2. HPC-AI Tech logo

    AI Engineer Intern

    2025-01 - 2025-08
    HPC-AI TechSingapore
    • Open-Sora 2.0 infrastructure: contributed optimization work for a former open-source SOTA text-to-video model, focused on faster generation and better output quality.
    • Video Ocean Backend (Go, Entgo, go-zero): built and maintained backend services for a commercial text-to-video platform serving 300,000 users.
    • Developed regular payment business modules and database maintenance SDKs.
    • Contributed to two major system reconstructions, enhancing scalability and stability.
    • Co-author (1 of 33) of Open-Sora 2.0 — a former open-source SOTA video generation model trained for $200k; 28,600+ GitHub stars, 91+ citations.
  3. Beijing Stats City Data Technology Co., Ltd. logo

    Full Stack Engineer

    2024-01 - 2024-09
    • ScholarHero (书卷侠): led a student startup team to develop an AI-powered educational application; reached 300 users.
    • Received recognition and funding support from the Communication University of China; attracted attention from top Chinese universities.
    • Published a first-author paper (FOKE, Springer CCIS, BDSC 2024); patents and software copyrights pending.
    • TextRAG + GraphRAG hybrid knowledge system (Python, Neo4j, ChromaDB).
  4. Lenovo (Beijing) Co. Ltd logo

    Python Develop Engineer Intern

    2023-06 - 2023-09
    • Knowledge and Training System: led a team of 3 interns to fine-tune large language models and develop Lenovo's internal AI training system (intelligent recommendation and virtual teaching) using Flask and Vue3.
    • Established the internal network for project deployment, configuring network equipment including NAS and computational resources.
    • Stable Diffusion Launcher: the precursor to the AI image generation software pre-installed on Lenovo AI PCs.
  5. Ipsos China logo

    Market Research Analysis Intern

    2022-07 - 2022-08
    Ipsos ChinaBeijing, China
    • Used Octopus crawler tool to collect product data for AIoT smart in-vehicle devices, extracted and analyzed questionnaire data using SPSS, and conducted industry surveys on current AIoT smart in-vehicle devices using Microsoft Excel.
    • Gained familiarity with market research processes, improved data collection, organization, and analysis skills, and received the Outstanding Project Award for the market research internship project.
    • Participated in immersive learning experiences at HSBC Bank, Deloitte China, Siemens (China) Co., Ltd., and Beijing DeepGlint Technology Co., Ltd., gaining in-depth understanding of banking, risk and financial consulting, auditing, management consulting, corporate structure, and AI technologies.
    • Gained comprehensive understanding of operations and culture in renowned companies; expanded skills in finance, management, auditing, and artificial intelligence; improved practical abilities in case analysis and business plan development.

Awards

Skills

Large language modelsMulti-agent systemsVideo generation modelsPrompt engineeringFoundation modelsBenchmark designGenerative engine evaluationMulti-agent orchestrationAgent runtime designMCP-like protocolsExecution systemsCoordination protocolsCapability invocationAI-native databasesProcedural memoryTextRAGGraphRAGNeo4jChromaDBKnowledge graphsRustGoPythonTypescriptAcademic writingFirst-author publicationsCCF-A publicationPatent filingResearch-to-product translationgRPCDistributed systemsProduction-grade systems engineeringInference accelerationBackend services