Addison Baum

Software Engineer (Founder)

Anigraph

Oct 2025 — Present

Built and deployed a large-scale anime discovery platform integrating collaborative filtering, content embeddings, and creative staff network analysis to power personalized recommendations and interactive exploration
Developed a hybrid recommendation engine blending collaborative filtering with content-based signals (tag similarity, staff networks, review-derived descriptive vectors), served via an in-memory embedding cache with hot-reload on model version changes
Trained an XSimGCL model on 252M user–anime interactions across 3.5M users; implemented embedding clustering for thematic discovery and a fold-in technique that embeds new users from their first few ratings, enabling real-time onboarding without retraining
Paired the embedding model with an XGBoost learning-to-rank reranker that combines collaborative-filtering signals with content and user-profile features to lift ranking quality beyond embedding similarity alone
Built a 4-phase NLP pipeline over the anime review corpus (spaCy POS tagging, weighted TF-IDF, LLM term curation, WordNet-based bipolar axis discovery) producing per-title descriptive vectors and axis scores used as content-based recommendation signals
Built a distributed Go ingestion pipeline pulling AniList GraphQL data through a 100-proxy pool with per-proxy rate limiting, exponential backoff, and resumable incremental updates; selectively recomputes recommendations and relationship graphs only for affected titles, keeping recomputation costs flat as the dataset grows
Designed a multi-store data architecture combining Neo4j for offline staff graph processing, PostgreSQL for transactional queries, and Elasticsearch for fuzzy full-text search across multiple indices
Built interactive D3.js force-directed graph visualizations for real-time exploration of staff collaboration networks with hierarchical filters and dynamic edge weighting

Software Engineer (Contract)

Talamel

Jan 2025 — Apr 2025

Built backend and real-time product features for a healthcare platform using Elixir, Phoenix LiveView, and PostgreSQL, contributing to application architecture and high-volume data retrieval paths
Improved PostgreSQL query performance by profiling query plans, reducing unnecessary allocations, and restructuring data-access patterns to reduce latency under load
Implemented dashboard and data-visualization workflows with Tailwind CSS, translating product requirements into usable frontend flows across the platform
Designed key components of the platform's UI in Figma, helping shape the overall visual and functional experience
Communicated technical decisions and delivery progress through written documentation while collaborating with product and engineering stakeholders in an Agile environment

Software Engineer (Contract)

Alcea Surrogacy

Jun 2024 — Aug 2024

Built an automated pipeline using Python and BeautifulSoup deployed as AWS Lambda functions to scrape external webpages updated weekly, detecting content changes across multiple public sources with configurable scheduling and error handling
Implemented PII filtering with regex pattern matching before sending content to the OpenAI API, generating concise weekly change summaries while maintaining data privacy
Replaced a manual review workflow, saving significant time while maintaining privacy and compliance with sensitive data handling practices

Varsity Overwatch Player / In-Game Leader

BU Gaming Club

Sep 2022 — May 2024

Led match strategy and in-game decision-making for a five-player team and drove segments of VOD review cycles while competing at a Top 250 individual rank (top 0.05% of players)
Elevated an unfunded program to compete with scholarship-offering teams, achieving the highest collegiate ranking in the program's eight-year history

Software Engineer Intern

SEO (Sponsors for Educational Opportunity)

May 2022 — Aug 2022

Developed a React web application with JavaScript and HTML/CSS featuring an interactive CSS Grid–based calendar visualization, implementing responsive UI components and modern functional component architecture with hooks
Integrated multiple RESTful APIs (OpenWeather, TimezoneDB) to fetch real-time UV index data and calculate timezone offsets, handling asynchronous data fetching and error states for a robust user experience
Built reusable React components including authentication forms, calendar grid, and navigation using Context API for global state management and React Router for client-side routing
Designed and implemented a Flask REST API backend with JWT authentication, MySQL database integration, and modular route architecture serving JSON endpoints for calendar data and user management
Collaborated in a team environment applying agile development practices, with input validation, error handling, and cross-browser compatible CSS for consistent user experience

Addison Baum

Featured Works

Anigraph

Facial Recognition Graph Search

AniList Metadata Scraper for Kodi

Experience

Software Engineer (Founder)

Software Engineer (Contract)

Software Engineer (Contract)

Varsity Overwatch Player / In-Game Leader

Software Engineer Intern

Education

Boston University

Certifications

Neo4j Certified Professional

Practical Network Penetration Tester

CompTIA Network+

Interests