Hi, I am James!

I'm a fullstack engineer. My favorite part of fullstack engineering is probably system design, as I love architecting end-to-end applications and taking ownership of projects. I am primarily interested in building LLM-based tools. My second favorite part about fullstack engineering is about making LLMs, both general-purposed and fined-tuned models, generate truly useful output.

These problems involve gathering high-quality inputs and contexts, giving LLMs clear instructions, and structuring algorithmic/agentic workflows. My hope is that the outputs produced from LLMs 1) opens up novel, valuable applications and/or 2) matches that of highly-trained, highly-skilled humans.

My current tech stack primarily is NextJS, TypeScript, TailwindCSS, though I especially am open to experimenting with different tools and languages (i.e. Scala, ProtoBuf) to introduce or strengthen correctness guarantees on my work. I often think about the social and economic impact of my applications. I want to build things that help people and society.

As a fullstack engineer, I am interested in learning and researching other disciplines that make fullstack applications performant, reliable, adaptable, and scalable. These disciplines include:

• Software Engineering Practices
• Machine Learning
• Natural Language Processing / LLMs
• Computer Vision
• Data Science
• Computer Security
• UI / UX and Graphic Design
• Theoretical Computer Science
• Algorithms and Data Structure Design
• Programming Languages (i.e. functional programming)

I enjoy startups for their dynamism. I love the freedom to set as high technical bars as I want for my work and pick problems that I find interesting. I love approaching problems from a very formal and theoretical manner. I especially am open to experimenting with different tools and languages (i.e. Scala) to introduce or strengthen correctness guarantees on my work.

In the past, I built multiple startup applications, where I performed all of the technical work. Among them were:

• Rx Jot, a LLM-based tool for doctors to write medical necessity and appeal letters for prior authorization
• PrepStar, a LLM-based tool to help students with practicing for AP exams
• TrialsLookup, a semantic search engine for doctors to find clinical trials for their patients

Quotes that I think about