Featured Works
Anigraph
Multifunctional anime discovery platform enabling users to discover new anime through shared creative staff relationships, tag/genre similarity analysis, and community-based clustering algorithms. Server-side rendered frontend using Nuxt.js/Vue.js for optimized SEO and sub-second initial page loads.
Visit Live Site →Key Features:
- → Architected Docker-based microservices stack orchestrating Nginx reverse proxy, Nuxt.js SSR application, Neo4j graph database with APOC and GDS extensions, Elasticsearch search engine, and Umami analytics
- → Leveraged graph database technology (Neo4j) for efficient relationship traversal and pattern matching, precomputed anime tag/genre cosine similarity matrices, and integrated Elasticsearch for full-text search with fuzzy matching
- → Implemented advanced graph algorithms using Neo4j Graph Data Science library including Louvain community detection for clustering anime creative teams and degree centrality for identifying influential staff members within clusters
- → Optimized database performance through strategic heap memory allocation, query result caching, and database indexing strategies, achieving 15ms average search response times and 30-60ms for complex multi-hop graph traversals in production
- → Developed interactive and responsive D3.js force-directed graph visualizations for exploring creative team connections. Created intuitive categorization through colored edges and nodes representing different relationship types and entity categories
- → Engineered complete ETL pipeline with parallelized web scraping from GraphQL API for data extraction, Python transformation scripts for data cleaning and normalization, and bulk loading into Neo4j and Elasticsearch. Processed thousands of anime entries with hundreds of thousands of relationship edges
Nuxt.js 3
Vue 3
TypeScript
Neo4j
Neo4j GDS
Elasticsearch
D3.js
Python
Docker
GraphQL
Click to expand
Facial Recognition Graph Search
OSINT tool designed for analyzing and visualizing relationships between images, supporting research and relationship mapping workflows. Enables recursive graph-based investigations through multi-face detection and dynamic graph expansion.
Key Achievements:
- → Reverse-engineered obfuscated API responses and cookie-based authentication flow, implementing hex-to-ASCII decoding and analyzing stateless session validation
- → Engineered multi-face detection pipeline using dlib and OpenCV to automatically identify and extract additional subjects from search results, enabling recursive graph-based investigations
- → Developed interactive graph visualization interface with vis.js, enabling real-time node manipulation, relationship mapping, and dynamic graph expansion for investigative workflows
- → Implemented dynamic Excel report generation with openpyxl, programmatically embedding base64-decoded images, hierarchical node relationships, and metadata into structured investigation reports
- → Built resilient proxy rotation system using Redis list operations (LMOVE) to cycle through datacenter proxies, preventing rate limiting and IP-based blocking during high-volume searches
Python
Flask
dlib
OpenCV
Redis
openpyxl
Vis.js
Click to expand