CURRICULUM VITAE of Dixing (Dex) Xu

Last updated October 1, 2026

Research Interests

Autonomous agents for machine learning engineering and research, including tree search over code, agent harness design, evaluation pipelines, and recursive self-improvement of research agents. Earlier work on distributed machine learning systems, privacy-preserving machine learning at the IoT edge, and blockchain systems.

Education

• 2015 - 2020 B.Sc. (Hons) in Information and Computing Science, Upper Second Class, dual award from the University of Liverpool, UK and Xi'an Jiaotong-Liverpool University, Suzhou, China. Gap year in 2019 for research at Nanyang Technological University and Zhejiang University.

Honors and Awards

• Dev Bounty Winner, Network School (first cohort), 2024.
• 2nd Prize, XJTLU Blockchain Technology Application Innovation & Entrepreneurship Challenge, 2020.
• Hyperledger Diversity Scholarship, Hyperledger Global Forum, Phoenix, AZ, 2019.
• CNCF Diversity Scholarship, KubeCon + CloudNativeCon China, Shanghai, 2018.
• Special Prize (US$10,000), Wanxiang Blockchain Hackathon by Qtum, awarded to 1 of more than 20 teams worldwide, 2018.
• 1st Prize (team), EOS Hackathon Hangzhou, 2018.
• 3rd Prize (individual), EOS Hackathon Hangzhou, 2018.
• 3rd Prize, DoraHacks x BCH Faith Hack, 2018.
• 2nd Prize, XJTLU & PNP AI Innovation Hackathon, 2018.
• IBM Student Innovation Lab Program Award, 2018.
• 1st Prize, Hack x FDU 2017, top 1 of more than 70 teams from universities nationwide, 2017.

Professional Experience

• Weco AI, London, UK
  Member of Technical StaffMay 2026 - Present
  • Core engineer and co-maintainer of AIDE, the tree-search ML engineering agent that OpenAI used for MLE-bench evaluations.
  • Built the autonomous research agent that set 7 leaderboard records in OpenAI's Parameter Golf (25 PRs filed; best 5-seed mean val BPB 1.0645).
  • Agent harness design and evaluation pipelines on MLE-bench and ALE-Bench; recursive self-improvement of research agents.
• Weco AI, Remote
  AI Research ConsultantOct 2024 - Apr 2026
  • Full-stack R&D on AIDE: web UI for the aideml app, maintenance of the open-source agent, and research on self-improving agents.
• Hex Trust, Remote
  Software Development EngineerApr 2022 - Aug 2024
  • Backend services for HexAdmin (Go) and the HTM markets platform (Rust, Go, Node.js); platform infrastructure with Terraform, Kubernetes and Nginx.
• Hyperledger, The Linux Foundation, Remote
  Maintainer, Hyperledger Fabric Python SDKSep 2018 - Jun 2024
  Mentor, Support Decentralized Governance for Smart Contracts in Fabric Python SDKMar 2021 - Nov 2021
  Mentor, Add Interactive Examples to Fabric Python SDKMar 2020 - Nov 2020
• Hong Kong Baptist University, Hong Kong
  Research AssistantOct 2020 - Nov 2021
  • Distributed machine learning systems and blockchain systems, with Prof. Xiaowen Chu.
• The Linux Foundation, Remote
  Software Development Engineer (Blockchain) Intern, Hyperledger CelloJun 2020 - Dec 2020
• Zhejiang University, Hangzhou, China
  Visiting Student ResearcherSep 2019 - Mar 2020
  • Machine learning and security, with Prof. Peng Cheng.
• Nanyang Technological University, Singapore
  Research AssistantMar 2019 - Aug 2019
  • Collaborative learning, IoT and privacy, with Prof. Rui Tan; led the privacy-preserving ML in IoT project.
• The Linux Foundation, Remote
  Software Development Engineer (Blockchain) Intern, Hyperledger Fabric Python SDKJun 2018 - Nov 2018
  • Mentored by Dr. Baohua Yang.
• Xi'an Jiaotong-Liverpool University, Suzhou, China
  Research AssistantNov 2016 - Aug 2018
  • Deep reinforcement learning for portfolio management, with Dr. Jinjun Liang.
  Summer Undergraduate Research FellowJun 2017 - Aug 2017
  • Style transfer with generative adversarial networks, with Prof. Steven Guan.

Publications

Refereed Journals

1. Lightweight and Unobtrusive Data Obfuscation at IoT Edge for Remote Inference, by D. Xu, M. Zheng, L. Jiang, C. Gu, R. Tan, P. Cheng. 2020, IEEE Internet of Things Journal (IEEE IoT-J). doi:10.1109/JIOT.2020.2983278

Refereed Workshops

2. Challenges of Privacy-Preserving Machine Learning in IoT, by M. Zheng, D. Xu, L. Jiang, C. Gu, R. Tan, P. Cheng. Pages 1-7, 2019, First International Workshop on Challenges in Artificial Intelligence and Machine Learning for Internet of Things (ACM AIChallengeIoT). doi:10.1145/3363347.3363357

Preprints

3. Recursive Self-Improvement of AI Research Agents, by D. Srikanth, B. Zhao, D. Xu, Y. Wu, Z. Jiang. 2026, arXiv:2609.26457. arxiv.org/abs/2609.26457
4. AIDE: AI-Driven Exploration in the Space of Code, by Z. Jiang, D. Schmidt, D. Srikanth, D. Xu, I. Kaplan, D. Jacenko, Y. Wu. 2025, arXiv:2502.13138. arxiv.org/abs/2502.13138
5. A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem, by Z. Jiang, D. Xu, J. Liang. 2017, arXiv:1706.10059. arxiv.org/abs/1706.10059

External Presentations

1. 2026, Hands-on AutoResearch: Cracking OpenAI's Parameter Golf, workshop with the Weco AI team, AI Engineer World's Fair, San Francisco.
2. 2024, Algorithmic Trading Workshop, Network School. Slides
3. 2019, Deep Learning for Power System Security Assessment. Slides
4. 2019, Introduction to Hyperledger Fabric. Slides

Open Source Software

• AIDE: tree-search agent for machine learning engineering (co-maintainer): github.com/WecoAI/aideml
• Hyperledger Fabric Python SDK (maintainer, release manager for v0.7): github.com/hyperledger/fabric-sdk-py
• seedance2-skill: agent skill for writing Seedance 2.0 video prompts, 4.1k stars: github.com/dexhunter/seedance2-skill
• Vibe Clock: usage tracking for Claude Code, Codex and OpenCode: github.com/dexhunter/vibe-clock
• Merged contributions to 58 open-source projects (180+ pull requests), including UK AI Security Institute's Inspect, openclaw, AutoGPT, goose, Qwen Code, Pydantic AI, ShinkaEvolve and OpenEvolve: github.com/dexhunter
• QtumSwap: atomic swaps between Qtum, BTC and QRC20 tokens (Wanxiang Blockchain Hackathon): github.com/PRIEWIENV/QtumSwap
• DSharing: decentralized house renting on blockchain and IoT (Hack x FDU 2017): github.com/dexhunter/hackxfdu-DSharing

Major Media Exposure

• OpenAI's Parameter Golf write-up featured a model-compression PR from my research agent, 2026: openai.com/index/what-parameter-golf-taught-us
• Meet the Hyperledger Summer 2018 Interns, Hyperledger blog, 2018: hyperledger.org/blog
• Interview on the Hyperledger internship (in Chinese), 2018: archive.is/5nyBm
• Coverage of the deep reinforcement learning portfolio management framework (in Chinese), 2017: archive.is/XKqY3
• Coverage of the Hack x FDU 2017 win (in Chinese), 2017: archive.is/FzYVh