Founding Engineer
I build AI-driven web applications and create innovative solutions at the intersection of machine learning and software engineering.
I'm a full-stack Ruby on Rails engineer focused on building internal tooling, API integrations, and simplifying complex systems. My expertise lies where web apps meet AI, creating experiences that are satisfyingly simple and intelligent with fast, scalable performance.
Currently, I'm a Founding Engineer at Luthor, where I build LLM observability tooling, OAuth integrations, and compliance systems. Previously, I co-founded Los Angeles AI Apps, delivering custom web applications and AI integrations for businesses and startups.
I've had the good fortune to work on a variety of projects in both machine learning and web development. This includes a link-in-bio tool to help users organize their events on social media; a newspaper personalized for you that uses AI to scour sources like Hacker News and Arxiv; and a chatbot that answers questions about AI, JavaScript, and Ruby.
In my free time, I enjoy blogging about making web development simple, indoor gardening, Brazilian jiu-jitsu, and Stanley Kubrick films.
A newspaper personalized for you. Add your favorite topics, and have AI scour sources such as Hacker News and Arxiv to find the most interesting and pressing stories of the day. A new issue is generated daily, always with fresh stories to read and enjoy.
With Amble, you can create a public event calendar and share your link on socials to turn followers into attendees. They can subscribe and instantly see updates in their favorite calendar app. Add one-click links for navigation, your website, Yelp, and more.
Do you love riddles? Have some time to kill? Then try your hand at some tricky riddles on the Riddler app! Have a whimsical gnome powered by a large language model give you hints to guide you in the right direction.
Have a question about artificial intelligence, JavaScript, or Ruby? Yapper is a chatbot that will answer your questions and help you discover new blogs by some of top minds in their respective fields. And it's free!
A film recommendation Q&A bot using semantic search and metadata filters. Database of ~10,000 popular films (1950-2023) with retrieval-augmented generation for intelligent recommendations.
Machine learning application for plant disease detection using object detection to localize and classify diseases. Built with Metaflow for orchestration and deployed on AWS SageMaker.
Segmented cars from their background using images from the Carvana Kaggle competition. Deep dive into U-Net architecture and image segmentation techniques.
Coded a mini-GPT model from scratch, training on Shakespeare text. Implementation of transformer architecture and attention mechanisms.
Designed the jumping mechanics of the Mini Cheetah robot as it maneuvered through obstacles. Applied control theory and optimization techniques.
Architected LLM observability dashboard instrumenting AI pipeline jobs with prompt/response capture and latency metrics. Built OAuth 2.0 integrations for Jira and Twitter/X APIs with JWT-secured webhooks and pessimistic locking. Redesigned document annotation pipeline, reducing complexity from O(n) fuzzy matching to O(1) range lookups. Shipped disclosure management and multi-language support systems for global compliance review.
Built and shipped Ruby on Rails applications with AI integrations; owned product/UX decisions and rapid iteration from prototype to production. Architected The Digital Press, a Rails 8 system combining LLM workflows with semantic search to process 1,000+ articles/day. Implemented responsive UIs with Stimulus/Turbo and production background job workflows.
Secured and managed a $3M federal grant; led cross-functional stakeholder coordination and clear technical communication at scale. Developed Python-based simulations using NumPy and Matplotlib to enhance engineering curriculum, introducing programming as a pedagogical tool for 150+ students per semester.
Taught four classes of regular and Advanced Placement Physics, two classes of Integrated Coordinated Science, and one Robotics class to approximately 75 students per semester. Adapted science lessons for Next Generation Science Standards, making them more interactive and student-focused.
New writing lives on my site. View all posts
A beginner-friendly guide to building your first AI-powered web application with minimal complexity. Learn the fundamentals of integrating AI APIs into web applications.
Explore how to leverage Ruby and Rails for AI application development. Practical examples and best practices for integrating machine learning models with Ruby on Rails.
Demystifying the AI technology stack from infrastructure to applications. A comprehensive guide to understanding how different layers of AI systems work together.
Can we really do RAG with a single-file architecture? Learn how to perform retrieval-augmented-generation using SQLite with the sqlite-vec extension and OpenAI API.
On the virtues of radical simplicity. Why SQLite is rapidly becoming a production-ready database for web applications and machine learning.
Adding evaluation, automated data pulling, and other improvements to take a film recommendation RAG app from development to production.
How to create a production-grade pipeline for object detection, bridging the gap between data science development and production deployment.
RAG + Filtering with Metadata = Great Movie Recommendations. Build a film recommendation system using self-querying retrieval.
Object detection using the power and simplicity of the KerasCV library. A comprehensive guide to building an end-to-end object detection pipeline.
How can you get a computer to distinguish between different types of objects in an image? A step-by-step guide to U-Net architecture and implementation.