Cupertino, CA · Pittsburgh, PA · nehemiah.newell@gmail.com · GitHub · RefChecker · LinkedIn
Education
Arizona State University — B.S., Computer Science / Software Engineering (2025)
Experience summary
Recent Computer Science grad with full-stack and systems software skills using Python, C++, Ruby, Swift and REST APIs. Projects include an AI-assisted reference validation tool, a file manager with embedded AI, and an iOS music delivery platform. Strong foundation in modular design and test-driven development, GenAI, and agile methods.
Core skills
- Languages: Python, C++, Java, Ruby, Swift, PHP, SQL, JavaScript
- Frameworks / web: React, Angular, Django, Rails, MVC patterns, REST APIs
- GenAI engineering: Tool orchestration, multi-turn planning, structured prompting
- Tools & infrastructure: PyCharm, Claude Code/CoWork, Gemini, OpenAI
Technical projects
TwinsMusic: iOS delivery platform (senior capstone)
Collaborated in a distributed three-person team using Agile standups to deliver iOS client functionality for an e-commerce music catalog:
- Designed and delivered a backend-driven UI in Swift, allowing dynamic control of navigation and page availability without requiring app store updates.
- Led development of authentication flows, user account state handling, and integration with Apple StoreKit.
SQL & web architectures
Built a multi-user SQL database (SQL Server) survey and testing system with a C# front end: comparison tests and surveys, editable in place, all test and user information stored server-side and accessed remotely, with update paths that let surveys grow or change without losing information or context. Developed full-stack applications including a Ruby on Rails forum with admin/user permission logic, a Django hashtag analytics tool for conversation tracking, and a PHP e-commerce storefront.
RefChecker: AI-assisted reference validation
Hallucinated references are the bane of academic paper reviews. RefChecker automates reference validation using Gemini/GROBID extraction and programmatic verification against authoritative sources (CrossRef, PubMed, and others). Engineered a modular plugin architecture with a Tkinter GUI to support individual and batch processing of papers. Built with Claude Code; integrates a personal git system for code and context management and embeds Gemini for extraction and validation.
Writer’s Assistant: LightRAG-powered knowledge base & wiki
A RAG system paired with a Karpathy-inspired wiki structure: it ingests complex documents and automatically generates interconnected wiki pages without losing context, tracking relationships between entities while maintaining an unaltered record.
- Implemented a LightRAG backend to extract and use graph-based relationships, automatically discovering and mapping connections across the wiki.
- Deployed AI agents to filter raw document ingestion, determining the initial selection and structure for the starting wiki.
- Built a chatbot that uses the wiki as its knowledge base.