Building with
|

Multi-agent system with interpretable communication

Research framework exploring how AI agents develop interpretable languages. 85% human readability vs 15% baseline. Open source with 6 stars.

Reinforcement learning for adaptive trading

Stock trading robot using RL with reward-action-correction framework. 76 stars, 15 forks. Focus on risk management over returns.

Pattern recognition with CNN/LSTM ensemble

Market analysis using deep learning for pattern recognition. 12 stars, 3 forks. Emphasizes research over deployment.

Collaborative agents for currency markets

Multi-agent RL for forex trading. Agents specialize in different strategies and coordinate positions. Early-stage research.

ML experiments and learning playground

Personal collection of ML experiments, utilities, and research explorations. Named "牛逼" as reminder to build skills.

Autonomous agent experiments

Framework for building AI agents with planning and reasoning. Exploring agent architectures and tool use.

Adaptive trading strategies

Exploring meta-learning for strategy adaptation in MetaTrader 5. Few-shot learning for market regime changes.

Microservices system for learning

Led team building AI-powered education platform. FastAPI microservices, BERT models, real-time analytics. Cannot share code (proprietary).

ML-powered travel recommendations

Co-founded travel tech startup. Built hybrid recommendation engine with collaborative filtering and LightGBM. Cannot share code.

Current startup project

Co-founded fintech startup building trading infrastructure. Backtesting engine, ML pipeline, AWS deployment. Work in progress.

BERT-based NLP system

Built Chinese text classification with fine-tuned BERT. Part of education platform work. Optimized inference latency.

ETL with Apache Airflow

Built ETL pipelines processing daily records from multiple sources. Part of education platform infrastructure.