About

Computer Science senior at the University of Arizona (3.85 GPA, Dean's List) with a focus on AI systems and software engineering. I build things at the intersection of language models and real products, from RAG-based GPT frameworks and LLM evaluation pipelines to founding backend systems for early-stage startups. Currently engineering AI infrastructure at UA × McGraw Hill while completing my degree in May 2026.

Jan 2026 – Present Active

AI Systems Engineer POC

University of Arizona × McGraw Hill  ·  Remote

  • Architecting a controlled, RAG-based GPT framework enabling instructors to create textbook-grounded, task-specific AI assistants.
  • Enforcing role-based and content-level guardrails to constrain retrieval and model behavior to approved instructional materials.
  • Designing source-aware response patterns that reference specific textbook chapters or sections to improve transparency and trust.
  • Optimizing retrieval design, chunking strategies, and evaluation criteria to improve response consistency and reduce hallucination risk.
RAG LLM Systems Prompt Engineering AWS Bedrock
Sep 2025 – Present Active

Computer Science Fellow

Handshake AI  ·  Remote

  • Building reproducible Docker-based test environments to enable automated evaluation of AI software engineering agents at scale.
  • Re-implementing bug fixes across unfamiliar codebases as ground-truth solutions, authoring fail-to-pass test cases and structured problem prompts for AI agent training.
  • Analyzing failure patterns in agent outputs to refine evaluation pipelines and improve benchmark reliability across iterative testing cycles.
AI Agent Benchmarking Docker LLM Evaluation Fail-to-Pass Testing
Oct 2025 – Jan 2026

Founding Engineer

SkillfullyAware (SAAQ) App  ·  Tucson

  • Designed scalable, API-driven backend services to process structured and unstructured user input into analytics pipelines.
  • Implemented RESTful services powering frontend features and internal system integrations.
  • Led MVP backend development across core services, data models, and application workflows under evolving product requirements.
  • Integrated external LLM APIs and vector-based retrieval components to support early-stage AI feature experimentation.
REST APIs Vector Databases LangChain MVP Development
Aug 2025 – Dec 2025

AI Solutions Developer – RAG Systems

Andrew Weil Center for Integrative Medicine  ·  Tucson

  • Built an AI-powered RAG chatbot for the CanHeal Cancer Toolkit, deployed on AWS microservices to streamline oncology resource access.
  • Conducted user research and iterative testing with medical stakeholders to ensure usability and HIPAA compliance.
  • Produced technical documentation and delivered client-ready demos for faculty and sponsors.
AWS RAG Microservices Agile
Sep 2023 – Aug 2025

GRC Automation Analyst

University of Arizona IT Services  ·  Tucson

  • Automated ServiceNow GRC workflows, cutting manual review time by 25% and improving policy audit traceability.
  • Developed SQL/PostgreSQL dashboards tracking compliance risk metrics used by three internal teams.
  • Delivered Agile-based feature updates under tight deadlines while maintaining Git version control integrity.
ServiceNow SQL Agile Git
2026

guardrail-auditor

LLM security auditing SaaS that red-teams prompt-injection, data leakage, and role-bypass vulnerabilities — built with Next.js 14, Prisma, and Tailwind CSS, deployed on Vercel.

TypeScript Next.js 14 Prisma AI Safety SaaS
2026

c-coroutine

Cooperative coroutine (fiber) library in C — hand-written x86-64 context switch, FIFO scheduler, and co_yield/co_await primitives with zero external dependencies.

C x86-64 Systems Programming Concurrency
2026

go-ratelimiter

Thread-safe rate limiting library in Go implementing token bucket, sliding window, and leaky bucket algorithms — zero dependencies, designed for use as middleware in backend services.

Go Rate Limiting Middleware Concurrent
2026

Library-Management

Java desktop library management app with book search, to-read lists, ratings, and a full JUnit test suite — built with Swing and a clean OOP/MVC design pattern.

Java Swing OOP/MVC JUnit
2026

go-lrucache

Thread-safe, generic LRU cache in Go with O(1) get/put and optional TTL-based eviction — designed for high-concurrency workloads with zero external dependencies.

Go Generics Data Structures Concurrent
Currently Focused On
  • AI system evaluation, reliability, and reducing hallucination in production RAG pipelines
  • Automation in risk, compliance, and enterprise workflow systems
  • Scalable backend architecture and cloud infrastructure design
Beyond Tech

When I'm not building or coding, I like pushing myself in other ways, whether through training, sports, or learning about geopolitics and human behavior. I'm naturally curious about how complex systems work, and that perspective shows up in the way I think about software and real-world problems.