Experience

GE HealthCare

• Built a modular Go-based API client designed for Terraform integration, featuring secure OAuth2 authentication and encrypted token caching through multiple configurable methods.
• Developed a custom Terraform provider integrating the API client to enable infrastructure-as-code provisioning of tenant organizations, supporting full CRUD operations, state management, and reusable modules.
• Collaborated in a cross-functional Agile Scrum team to gather requirements, design, and implement a solution that automated tenant onboarding, saving 4 days of manual work per new organization.

Campus Life Office, Dickinson College

• Built a web application using React, JavaScript, and CSS for Resident Advisor housing assignment, saving 8+ hours of office time and resources; deployed it using Vercel.
• Led residential areas with a total of 200+ residents, fostering an inclusive and supportive community, and organized 15+ community-building events.
• Awarded the Academic Excellence Award for outstanding academic performance and rehired for a third year in a row.

Academic Technology Department, Dickinson College

• Developed an Arabic AI chatbot on a UNIX/Linux environment using Python, OpenAI API, and locally hosted LLMs via Ollama to assist students in learning Arabic, increasing engagement by 70%.
• Engineered system prompts to guide LLM behavior and integrated LangChain for chat memory and input validation, reducing prompt injection risk by 60% and improving contextual consistency by 50%.
• Enabled professor-driven customization of chatbot behavior via the frontend and implemented SMTP integration for students to email formatted chat transcripts.

84 Lumber

• Participated in a project to modernize and document 50k+ lines of legacy CBASIC code to Python/Java by developing an NLP pipeline for automated code conversion and documentation.
• Implemented an Abstract Syntax Tree (AST) pipeline using data structures and algorithms (BST, DFS, BFS) and ANTLR4 to create a CBASIC grammar definition program while maintaining hierarchical code structure.
• Researched Hugging Face Transformers and pre-trained language models like BERT and GPT to evaluate the efficiency of AI-powered code documentation and modernization.