• Utilized Go to engineer a limited personal information API at the above URL
• Containerized app using Docker and deployed using Google Cloud Run with custom domain
• Utilized Svelte to build documentation page to explain API endpoints to an empty audience
• Wrote multiple unit tests to verify accuracy of information fetched from API endpoint

• Built a CLI tool using Go to help developers build CLI tools with OpenAI and Anthropic
• Utilized Cobra to create CLI tool with subcommands and flags for optimal user experience
• Created local storage solution for user API keys and configuration options
• Published Krait to Github where anyone with Go installed can download and use it

• Developed defense copilot to guide decision making based on standard operating procedures
• Built frontend with satellite imagery and chat interface using Next.js, shadcn, and Tailwind
• Engineered Flask backend to handle image and chat requests using Supabase and Anthropic
• Prototyped RAG agent using Adalflow and Supabase to make decisions based on military doctrine
• Competed at Y Combinator hackathon, ultimately featured by Adalflow as a prime use case

• Developed an application to offset carbon emissions from LLM calls used in AI apps
• Created a custom Python library (call-to-change) to track LLM calls through function wrappers
• Utilized MongoDB to track text generation and image generation calls in a cloud database
• Built a secure emissions dashboard interface using NextJS, Tailwind, and Clerk

• Created a chatbot to help students based on 3,898 UCSD undergraduate course descriptions
• Used Supabase to store and retrieve vector embeddings of course descriptions
• Utilized Langchain.js to chain prompts, fetch Supabase data, and interact with the OpenAI API
• Created an interactive interface using Next.js, shadcn, and Tailwind

• Developed a focus tracker that punishes users through audio when a phone is detected
• Wrote Python script to automatically capture photos and save them locally using OpenCV
• Collected, processed, and labelled object detection boxes for 827 images using Roboflow
• Incorporated YOLOv5 to train the model and detect phones in webcam feed
• Achieved mAP of 98.4%, precision of 96.9%, and recall of 96.4%

• Developed a computer vision app that analyzes food images and provides caloric information
• Utilized the OpenAI API to generate recipes based on detected ingredients in image
• Engineered computer vision pipeline to fetch a list of detected fruits and vegetables
• Won 1st place in a computer vision hackathon with team of 3

• Developed an Alexa skill that tells users the time remaining until the next F1 Grand Prix
• Created relevant intents, sample utterances, and slots through Alexa Developer Console
• Used Python to parse F1 schedule data fetched from the Ergast F1 API

• Developed a web app that finds anti-podal location (location on opposite side of the planet)
• Used Streamlit to create map that displays both user location and anti-podal location
• Used reverse geocode library and GeoPy to indentify nearest countries when in water
• Utilized Streamlit Share and Github to deploy web app