hunliang

Built a personalized crypto news aggregation engine using Go, PostgreSQL, Redis, and Next.js that prioritizes relevant content based on user watchlists, tagging key sentiment shifts and market trends through RESTful APIs and WebSocket connections. Implemented advanced topic modeling algorithms (LDA, BERTopic) with Python, scikit-learn, and Hugging Face Transformers, to extract underlying themes from crypto news sources, enabling users to discover hidden market narratives and interconnected industry developments. Developed predictive time-series forecasting frameworks like NeuralProphet that track sentiment momentum patterns and their corresponding price trajectories, achieving 78% directional accuracy on major tokens.

Integrated microservice architecture using Docker, AWS EKS, and Terraform, reducing deployment time by ~65% and improving system scalability to handle increased traffic loads. Deployed Big Bang 2.0 Helm chart for deploying DoD-hardened packages into a Kubernetes cluster, streamlining security compliance workflows and reducing manual configuration errors. Implemented and documented disaster recovery plans for the Kubernetes cluster, achieving < 4-hour recovery time objectives and remediating 95% of critical security vulnerabilities within 48-hour SLAs.

Researched in the Early Detection of Proliferation Activities team to research Event and Anomaly Detection data pipeline using machine learning and natural language processing, particularly phrase vectorization and embedding generation, using Python.

Implemented, documented, and benchmarked new techniques for integrating scientific visualizations with HTML5 and X3D in ParaView and Castle Game Engine through Python automation, improving rendering performance by ~40% and reducing load times for complex 3D medical datasets. Submitted a short paper titled “Enhancing Brain Flow Visualization with Automated 3D Data Processing: A Study on DCE-MRI Data from Mice with Tumors” to the ACM Web3D conference. Won 2023 VTURCS 3rd Place Research & 3rd Place Industry Awards within the Department of Computer Science at Virginia Tech annual research symposium.

Hosted weekly labs, lectures and office hours for several courses including Python II (CS1064, CS2064), Computer Organization I (CS2114), Java I and II (CS2114)