CV

CV of Junming Huang — AI researcher at DUAL Research Group, University of Sydney.

Contact Information

Name Junming Huang
Professional Title AI Researcher & Engineer
Email jhua0805@uni.sydney.edu.au

Professional Summary

Versatile AI researcher and engineer with a track record of rapid mastery across disciplines, transitioning from Cybersecurity to frontier AI. Currently exploring multiple frontier topics at the DUAL Research Group, specializing in Distributed AI and cross-domain R&D. Combines high-intensity engineering execution with strong leadership and communication skills. Dedicated to bridging theoretical innovation and robust implementation in fast-paced, demanding R&D environments.

Experience

  • 2025 - present

    Sydney, Australia

    Research Member
    DUAL Research Group, University of Sydney
    Working on Distributed AI, LLM Evolution, and Neural Dynamics research.
    • Lab Infrastructure: Server maintainer and computational resource scheduler; managing Linux-based environments and coordinating GPU allocation for secure research operations
    • Co-authored 2 papers (2nd and 3rd author) published at the 2025 6th CAIT
    • 1st author paper under review at IEEE Internet of Things Journal (resource-efficient AI)
    • 2nd author paper under review at IEEE Networking Letters (PEFT in Federated Learning)
  • 2021 - 2024

    Guangzhou, China

    MLPS Assessor
    Guangzhou Jingyuan Security Technology Co., Ltd
    Conducted Information System Security assessments and compliance audits under MLPS 2.0.
    • Specialized in security hardening for Windows/Linux systems and regulatory coordination
    • Earned National Junior MLPS Assessor certification within one year
    • Managed multiple concurrent security projects with cross-departmental and client-regulatory coordination

Education

  • 2024 - present

    Sydney, Australia

    Master of Computer Science
    University of Sydney
    Computer Science
    • Expected graduation: November 2026
    • WAM: 85/100 (High Distinction)
  • 2017 - 2021

    Ningbo, China

    Bachelor of Materials Physics
    Ningbo University of Technology
    Materials Physics
    • Ranked 3rd in major
    • Systematic laboratory research experience in material science: synthesis, characterization, and data analysis

Projects

  • CV-based Technology for Endangered Species Conservation

    Joint research project with BirdLife Australia exploring the feasibility of individual identification of Regent Honeyeaters (fewer than 200 remaining in the wild) using Computer Vision. Developing innovative, low-cost technical methods to overcome the limitations of traditional tracking.

    • University of Sydney collaboration with BirdLife Australia
    • Multi-state monitoring of critically endangered birds
  • End-to-End LLM Training and Evolution

    Self-driven project re-implementing the MiniMind LLM workflow, extending from pre-training to SFT, LoRA, and RLHF-based alignment. Released full model weights on Hugging Face (J1mm7/mini_lm).

    • Hugging Face release: J1mm7/mini_lm
    • Expanding to multi-modal capabilities (MiniMind-V)
    • Researching from-scratch implementations (Karpathy’s nanochat)
  • Causal Analysis of Climate Data with Machine Learning

    Studied traditional algorithms (SVM, RF) and neural architectures (RNN/LSTM/CNN) using PyTorch/TensorFlow for climate modeling. Applied Mutual Information and Transfer Entropy to CMIP climate data, identifying significant lagged causal relationships between surface temperature and long-wave radiation.

    • Lawrence Berkeley National Laboratory (LBNL), USA (online)
    • Applied Mutual Information and Transfer Entropy to CMIP climate data

Skills

AI Research Stack (Expert): PyTorch, Flower, DeepSpeed, Federated Learning, LLM Fine-tuning, LoRA, RLHF
Programming (Expert): Python, Linux Systems Administration, GPU Cluster Management
AI Engineering (Expert): CLIs, MCP, Agentic workflows, Complex project planning

Languages

Cantonese : Native speaker
Mandarin : Native speaker
English : C1 / Professional

Interests

Distributed AI: Federated Learning, Edge Deployment, Resource-Efficient AI
LLM Evolution: Reinforcement Learning, Fine-tuning, Inference Optimization, PEFT
Neural Dynamics: Neural ODEs, Flow Matching, Dynamic Systems

Awards

  • 2025
    INFO5990 Best Group & Best individual
    INFO5990 Teaching Team, University of Sydney

    Led a group project that won ‘‘Best in Semester’’ in an IT practice unit.