🎓 UG4 Artificial Intelligence & Computer Science @ University of Edinburgh
🧪 Ex-Intern @ Tencent Quantum Lab / Ex-Research Assistant @ IADS-UoE
💡 Focus: Intelligent Systems · Multi-Agent RL · ML Infrastructure · Scientific/Quantum Lab Software
📍 Edinburgh, UK | ✉️ ricardoshuoyu@gmail.com
- 2025 — Technology patent pending (Tencent Quantum Lab)
Publish–subscribe data distribution architecture for large-scale parallel experiments; client subscription to custom data hierarchies improving consistency and efficiency.
Preprints, reports, and technical notes.
| Year | Title | Venue / Type | Link |
|---|---|---|---|
| 2025 | Multi-Agent Reinforcement Learning Environment Design for LLM-Agent Simulation | UG4 Thesis (in progress) | – |
| 2024 | HAR on Edge Devices: Accuracy–Latency–Energy Trade-off | Course Report | |
| 2024 | Latency vs Distance in Edge-AI Pipelines | Short Technical Note | |
| — | (Future publications to be added) | – | – |
May – Sep 2025 · Shenzhen, China / Edinburgh (Hybrid)
- Designed a publish–subscribe data distribution architecture enabling hierarchical client subscriptions for consistent multi-device data sync. (Technology patent pending)
- Refactored LabRAD interfaces, replacing recursive calls with batch-processing algorithms, achieving 220× speedup (14m40s → 4s) in large registry operations.
- Built an automated confocal microscopy pipeline (data validation → conversion → CVM upload → PostgreSQL update → WeChat Bot notification).
- Deployed in production within Tencent Quantum Lab’s experimental infrastructure.
Stack: Python · Scala · PostgreSQL · LabRAD · Twisted · PyQt5 · Docker · REST
Jun – Aug 2024 · Edinburgh, UK
- Reproduced and extended the LineVul vulnerability detection framework.
- Conducted ablation studies across multiple ML models (BoW, Random Forest, CNN) achieving +15% F1-score improvement.
- Used Joern for static analysis of multi-base-unit (MBU) vulnerabilities; identified code pattern bottlenecks and proposed mitigation strategies.
- Authored a technical report summarizing results and recommended dataset refinements.
Stack: Python · PyTorch · Joern · scikit-learn
Supervisor: Dr. Wenda Zhou, University of Edinburgh (2025–26)**
A 2D Minecraft-inspired environment designed to study collaboration between RL and LLM-based agents.
Highlights:
- Implemented hierarchical and memory-augmented agents with prioritized replay and intrinsic motivation.
- Developed language-conditioned interfaces enabling LLM-driven subgoal decomposition and coordination.
- Built modular environment and benchmarking suite with Gymnasium + PyTorch + Pygame.
- Investigating multi-agent curriculum design and emergent communication.
Objective: To bridge symbolic reasoning from LLMs and adaptive behaviour from RL agents.
Course Project · IoT Systems & Data Science at Scale
- Designed an Edge → Cloud HAR pipeline using TensorFlow Lite, with adaptive noise filtering.
- Deployed quantized models to MCU; profiled latency–energy–accuracy trade-offs.
- Built data preprocessing workflow in Python + Spark, achieving 94% accuracy on 11 activity classes.
Course Project · Natural Language Processing (Feb–Mar 2025)
- Implemented Seq2Seq (GRU) model with Bahdanau attention for English–German translation.
- Visualized alignment maps confirming effective attention focusing on relevant source tokens.
Team Lead · Java REST API (Jan 2025)
- Built a RESTful backend for order management and path planning with no-fly zone constraints.
- Added GeoJSON export for visualization and Docker packaging for deployment.
Feb 2024
- Modeled game momentum using LSTM on 2019 Australian Open tennis data.
- Generalized framework to other two-player competitive settings.
Team Lead · Course Project (Oct 2023)
- Implemented TLB + Page Table simulation handling page faults and swapping.
- Achieved 147/150 final score; verified consistency with real OS behavior.
Team Project (Oct 2023)
- Cleaned and visualized 2016 mortality/admissions data using Matplotlib and Plotly.
- Conducted demographic-level comparisons by region and age group.
Machine Learning • Reasoning and Agents • Compiling Techniques • Computer Security • Mobile Robotics •
Natural Language Processing • System Design Project • Algorithms & Data Structures •
Discrete Mathematics & Probability • Computer Systems • Software Engineering
Languages: Python · Java · C · C++ · Scala · SQL
Frameworks/Tools: Spring Boot · REST · Docker · Git · PyQt5 · PostgreSQL
Domains: ML Systems · Scientific/Quantum Lab Software · Data Engineering · Evaluation & Benchmarking
