中文 | English | Français | 日本語 | Español | Deutsch
Zhili He

Zhili He

何致力 (Zhili He)

Researcher in Distributed Systems, Blockchain, and Artificial Intelligence

M.Sc. in Blockchain, Nanyang Technological University

B.Sc. in Digital Media Technology, University of Electronic Science and Technology of China (2025)

Email: HEZH0014@e.ntu.edu.sg | GitHub | Resume (Chinese) | Resume (English)

Biography

Currently pursuing a Master's in Blockchain at Nanyang Technological University, with a Bachelor's in Digital Media Technology from UESTC (2025). As a researcher in distributed systems, blockchain, and AI, I focus on scalability, security, and efficiency challenges through consensus protocols, zero-knowledge proofs, and AI-optimized systems. During undergrad, I gained hands-on experience in computer vision and NLP, contributed to industry-academia projects, and earned national scholarships and competition awards.

Research Interests

  • Blockchain and Distributed Systems: Exploring consensus algorithms, protocol design, smart contract security, and cross-chain integration for decentralized infrastructure.
  • Artificial Intelligence and Deep Learning: Advancing computer vision, medical image analysis, NLP, and multi-agent systems.
  • AI + Blockchain: Developing privacy-preserving AI, trustworthy ML, and secure architectures for blockchain-AI convergence.

Publications and Projects

Academic Papers

  • Z. He. (2024). Pneumonia image classification using convolutional neural network. Applied and Computational Engineering, 67, 255-266. DOI: 10.23919/ACE59475.2024.10283939 [Paper]

Research Projects

Hackathon_TornPrivacy

A hackathon project focused on enhancing and optimizing Tornado Cash's privacy protection mechanisms. This project delves into the application of zero-knowledge proofs and mixing technologies, achieving significant improvements in transaction privacy, opening new paths for blockchain privacy protection. The project overcame the complexity of privacy technology implementation and performance bottlenecks through innovative protocol design and optimization algorithms, ensuring a balance between privacy protection and efficiency. It also included security audits and user testing, contributing key technologies to the privacy coin ecosystem.

Tech Stack: Solidity, Zero-Knowledge Proofs, Mixing Technology, Blockchain Privacy, Protocol Design, Security Audit.

[Code]

AI-Powered Gastrointestinal Endoscopy Image Diagnosis System (2024-2025)

In the field of medical imaging diagnosis, a multi-branch vision model based on Mamba architecture achieved an outstanding test accuracy of 87.25% on the Kvasir dataset and successfully delivered a complete diagnostic system prototype. This achievement significantly improves the efficiency and accuracy of gastrointestinal disease diagnosis, bringing revolutionary changes to clinical medical practice. The project deeply researched the complexity of medical images and model generalization capabilities, overcoming challenges of limited datasets and class imbalance through multi-branch design and data augmentation techniques, opening new paths for AI applications in medicine. The system also includes model interpretability analysis and clinical validation processes, ensuring the credibility and practicality of diagnostic results.

Tech Stack: Python, PyTorch, Mamba, Medical Image Analysis, Computer Vision, Data Augmentation, Model Evaluation, Clinical Validation, Model Interpretation, Image Preprocessing.

[Code]

Distributed_Facility_Booking_System

Innovative design of a distributed facility booking system, implementing multi-node collaborative resource allocation and conflict resolution mechanisms. This project ensures high availability and fairness of facility bookings through distributed algorithms, providing technical support for the sharing economy. The project solved challenges in consistency and concurrency control in distributed systems, achieving stable operation and scalability through consensus algorithms and load balancing. The system also includes user interfaces and real-time notifications, providing a complete booking experience and setting benchmarks for distributed application development.

Tech Stack: Python, Distributed Systems, Consensus Algorithms, Concurrency Control, Load Balancing, User Interface, Real-time Notifications.

[Code]

Crypto Price Forecasting (crypto_forecast)

In the field of cryptocurrency price forecasting, the LightGBM time series modeling system achieved competition-level excellence through innovative "prediction reversal" strategies and training start date search mechanisms, with a final evaluation score of 0.08042. This breakthrough not only demonstrates advanced applications of cutting-edge technology in financial time series prediction but also provides reliable scientific basis for quantitative trading strategies. Facing challenges of high data volatility and complex non-linear market features, the project achieved significant improvements in prediction accuracy through meticulous feature engineering and model tuning, providing powerful tools for risk management in the cryptocurrency market. The project also includes advanced LGBM tuning, ensemble model fusion, and submission optimization strategies, covering the complete process from data preprocessing to model deployment.

Tech Stack: Python, LightGBM, Feature Engineering, Model Tuning, Time Series Analysis, Data Preprocessing, Cross-Validation, Ensemble Learning, Model Fusion, Submission Optimization.

[Code]

Singapore Major Bet (Singapore-Major-Bet)

A decentralized event betting platform project based on Ethereum smart contracts and Next.js full-stack architecture, implementing transparent on-chain betting mechanisms, real-time event listening responses, and dynamic visualization of prize pools. The platform not only integrates real-time leaderboards and wallet interactions but also deeply reveals the disruptive potential of blockchain technology in the gambling and entertainment industries. Overcoming challenges in smart contract security and user experience, the project ensures stable platform operation and user satisfaction through carefully designed contract logic and front-end optimization, setting benchmarks for commercialization exploration of Web3 applications. The platform also supports multi-chain deployment and smart contract auditing, demonstrating professional standards in blockchain development.

Tech Stack: TypeScript, Next.js, Flask, Web3.py, Smart Contracts, Ethereum, MetaMask Integration, Real-time Data Processing, On-chain Interaction, Contract Auditing.

[Code]

Kaia Projects

A collection of innovative projects in the Kaia blockchain ecosystem, including the BuyMeACoffee decentralized tipping platform, FX-linked stablecoins solutions, and kaia-scaffold-example development scaffolding. These projects explore the application potential of blockchain in finance and development tools, promoting the construction of the Kaia network ecosystem. The collection showcases the complete development process from proof-of-concept to practical tools, providing valuable references and resources for blockchain developers. BuyMeACoffee implements on-chain tipping mechanisms, the stablecoin project explores decentralized financial solutions linked to fiat currencies, and the scaffold simplifies the development process of Kaia applications.

Tech Stack: Solidity, Web3, Kaia Blockchain, Smart Contracts, Decentralized Applications, Stablecoin Design, Development Tools, On-chain Tipping, Financial Protocols.

[Code]

Depeg-Analysis

A deep analysis project of stablecoin depegging phenomena in decentralized finance, systematically studying the causes and impacts of stablecoin price fluctuations in the DeFi ecosystem. This project quantifies the triggering factors and market reactions of depegging events through data-driven methods, providing scientific basis for risk management. Overcoming challenges in blockchain data acquisition and multivariate analysis, the project achieved improved accuracy in depegging prediction through advanced statistical models and machine learning techniques, providing decision support for investors and protocol developers. The project also explores mechanism designs to prevent depegging, promoting the stable development of the DeFi ecosystem.

Tech Stack: Python, Data Analysis, Machine Learning, Blockchain Data, Statistical Modeling, Risk Assessment, DeFi Analysis.

[Code]

Enterprise Intelligent Dialogue System and Multi-Agent Collaboration Platform (2023-2024)

In the e-commerce field, the engineering practice of dialogue systems successfully deployed 30+ intelligent agents through data collection and cleaning, model fine-tuning, and large-scale deployment, improving dialogue accuracy to 91%. This project not only demonstrates the industrialization application of large-scale language models but also advances the frontiers of multi-agent collaboration systems. Facing the diversity and real-time requirements of e-commerce scenarios, the project adopted advanced RASA frameworks and LLM fine-tuning technologies, achieving seamless transitions from prototypes to production, providing replicable success cases for enterprise-level AI solutions. The project also covers performance monitoring, A/B testing, and user feedback integration, ensuring continuous optimization and user satisfaction.

Tech Stack: Python, RASA, LLM, Production ML, Natural Language Processing, Data Pipelines, Model Deployment, Performance Monitoring, A/B Testing, User Feedback.

RSA-Crypto

A deep implementation and optimization project of the RSA encryption algorithm, exploring the core mechanisms and practical applications of public-key encryption technology. This project not only implemented efficient encryption and decryption processes for RSA algorithms but also significantly improved the speed and security of encryption operations through mathematical optimization and performance tuning. The project addressed challenges in large number operations and key management, providing a solid theoretical and practical foundation for the field of information security. The project also includes complete implementations of key generation, distribution, and verification mechanisms, laying technical safeguards for modern encrypted communications.

Tech Stack: Python, Mathematical Algorithms, Large Number Operations, Encryption/Decryption, Key Management, Performance Optimization, Information Security.

[Code]

Buffotte

A masterpiece of intelligent analysis platform for CS2 item market, built with high-performance architecture using FastAPI and Vue3, achieving comprehensive integration of real-time candlestick charts, AI-driven price prediction, and market insights. The platform supports containerized deployment and automated pipelines, setting new industry benchmarks for game economy analysis. Overcoming challenges in complex game data acquisition and real-time prediction models, the project provides comprehensive market analysis tools through integrated LightGBM and front-end visualization, helping players and investors make informed decisions. The platform also includes personalized user recommendations and market trend analysis, enhancing user experience and decision support.

Tech Stack: Python, FastAPI, Vue.js, LightGBM, Docker, CI/CD, Data Visualization, API Integration, User Recommendations, Trend Analysis.

[Code]

buff-tracker

An API service designed for CS2 item price tracking, developing high-concurrency interfaces supporting single price queries, batch processing, and fuzzy searches, combined with Docker deployment and quota management, perfectly adapting to large-scale read-write demands. This system embodies best practices in API design and database optimization, providing solid guarantees for digital asset management. Addressing the rapid changes in the game market and diverse user needs, the project implemented efficient data indexing and caching mechanisms, ensuring service stability and response speed. The system also supports real-time price updates and historical data queries, meeting the needs of different user groups.

Tech Stack: FastAPI, Python, SQLite, Docker, API Design, Database Optimization, Concurrency Handling, Load Balancing, Data Caching, Real-time Updates.

[Code]

Twikk

A prototype platform exploring Web3 social scenarios, integrating MetaMask login and on-chain interactions, constructing an innovative model for decentralized social networks. This project deeply analyzes the disruptive application prospects of blockchain technology in the social media field, laying the technical foundation for future digital society. Facing the complexity of decentralized applications and user adoption challenges, the project demonstrates the possibilities of decentralized social networks through simple prototype design and Web3 integration, promoting the development of the Web3 ecosystem. The platform also explores NFT integration and decentralized identity verification, providing experimental insights into the future forms of social media.

Tech Stack: EJS, Node.js, Web3, Blockchain, Decentralized Applications, MetaMask, On-chain Interaction, User Authentication, NFT Integration, Identity Verification.

[Code]

hezhili.online

A personal academic and professional achievement showcase platform, built with modern web technologies to create a complete system including backend APIs and frontend interfaces. As an extension of the digital business card, this website comprehensively presents outstanding contributions in technological innovation and academic research. The project emphasizes the importance of personal brand building and online presentation, ensuring high-quality experiences on different devices through responsive design and content management. The website also integrates blog functionality and project showcases, providing comprehensive personal profile management solutions.

Tech Stack: Node.js, Express, HTML/CSS/JS, Web Development, Responsive Design, Content Management, SEO Optimization, Blog System, Project Showcase.

[Code]

Technical Skills

Programming Languages: Python (Expert), C/C++, C#, JavaScript/TypeScript, Go, Solidity, SQL

Deep Learning Frameworks: PyTorch, TensorFlow, Scikit-learn, Keras

Web Development: Vue.js, React, Node.js, Express, Nginx, HTML/CSS/JS

Blockchain Technologies: Ethereum, Smart Contracts, Web3.js, Solidity

Tools and Platforms: Git, Docker, Linux, AWS, Unity 3D, Jupyter Notebook

Others: Data Analysis, Machine Learning, Computer Vision, Natural Language Processing, Distributed Systems

Academic and Professional Experience

  • Master's Student - Nanyang Technological University, Blockchain Program. Currently focusing on cutting-edge research in the intersection of blockchain and artificial intelligence, committed to breaking through technical bottlenecks and driving industry innovation.
  • Bachelor's Degree (2025) - University of Electronic Science and Technology of China, Digital Media Technology. Completed studies with excellent grades, actively participating in multiple high-level research projects during this period, demonstrating outstanding academic abilities and innovative thinking.
  • Research Assistant (2023-2024) - Enterprise Intelligent Dialogue System Development Project. As a core member, improved dialogue accuracy to 91% through precise data processing and model optimization, setting benchmarks for large-scale AI applications.
  • Project Developer (2024-2025) - AI-Powered Gastrointestinal Endoscopy Image Diagnosis System. The deep learning model led achieved 87.25% diagnostic accuracy on the Kvasir dataset, bringing revolutionary breakthroughs to the field of medical imaging analysis.
  • Competition Participant - Provincial First Prize in National Undergraduate Mathematical Modeling Competition, National Third Prize in National Undergraduate Market Research and Analysis Competition. Excelled in fierce competition with rigorous logical thinking and data analysis abilities, showcasing outstanding performance in academic competitions.

Honors and Awards

  • 2024: National Third Prize / Provincial First Prize in National Undergraduate Market Research and Analysis Competition
  • 2023: "Model Student Scholarship" at University of Electronic Science and Technology of China * 2
  • 2022: Provincial First Prize in National Undergraduate Mathematical Modeling Competition