I recently had the privilege of participating in a hackathon centered on developing innovative solutions for the fine spirits industry, in partnership with the BAXUS marketplace. I tackled three distinct tracks, each designed to solve a unique challenge for whisky and wine enthusiasts. It was a challenging but incredibly rewarding experience, and I'm thrilled to share a detailed look at the projects I built.

Project BOB (BAXUS On-chain Bot)

This project was a deep dive into the world of artificial intelligence and personalization. My goal was to create an AI agent, affectionately named "Bob," to serve as a personal whisky sommelier. The core of this project was to analyze a user's existing collection within the BAXUS ecosystem and use that data to provide intelligent, personalized bottle recommendations.

Bob's functionality was designed around two key components:

Collection Analysis: The agent was tasked with parsing a user's bar data to identify patterns and preferences, such as their favorite regions, styles, and price points.
Recommendation Engine: Based on this analysis, Bob would then suggest new bottles. The recommendations were not just random; they included bottles within similar price ranges, those with similar flavor profiles, and even complementary bottles to help diversify a collection.

This project was a fantastic opportunity to explore how machine learning can enhance the user experience, transforming a digital collection into a dynamic and interactive personal tool.

Whisky Goggles

For this track, I focused on bridging the gap between the physical and digital worlds of spirits. "Whisky Goggles" is a computer vision system designed to be a practical tool for collectors on the go. My goal was to develop a system that could accurately identify whisky bottles by simply scanning their labels with a camera.

The system's core features included:

Label Recognition: I implemented a computer vision model capable of processing images of whisky bottle labels, extracting key visual features for identification.
Bottle Identification: The system was built to match the identified labels to the extensive BAXUS database of over 500 bottles. It was designed to achieve high accuracy even with variations in lighting, angles, or partial labels, and to provide a confidence score for each match.

This project was all about creating a reliable tool that could work seamlessly in real-world conditions, like a dimly lit liquor store, to help users quickly identify bottles and access relevant information.

The Honey Barrel

The final track I tackled was the "Honey Barrel," a solution aimed at helping enthusiasts find the best deals. This project was developed as a Chrome browser extension that automates the process of price comparison.

The extension's main features included:

Bottle Scraping: The tool was designed to automatically identify and scrape bottle information from various e-commerce and retail websites.
Price Comparison: It then cross-referenced this scraped information with the BAXUS marketplace API to compare prices, calculate potential savings for the user, and even account for factors like shipping costs.
Seamless User Experience: The results were displayed in a clean, non-intrusive popup, with clear visual indicators of potential savings and direct links to the BAXUS marketplace.

The Honey Barrel was all about creating a seamless experience that saves collectors time and money, and it served as an excellent case study in API integration and browser extension development.

A Unifying Application: Baxus-Frontend

My projects were all brought together under a unifying application, the Baxus-Frontend. This front-end application served as the user interface, allowing for a cohesive demonstration of each project's functionality. This was not a separate hackathon track, but a critical part of the overall submission, as it showcased how all three independent solutions could work together to create a powerful and integrated user experience.

Technical Highlights

Each project showcased different technical skills:

AI/ML: Natural language processing and recommendation algorithms for Project BOB
Computer Vision: Image processing and object recognition for Whisky Goggles
Browser Extensions: Chrome extension development and web scraping for Honey Barrel
Full-Stack Development: React-based frontend application for the unified interface

The Hackathon Experience

The BAXATHON was an inspiring event that allowed me to push the boundaries of what's possible in the spirits space. Working on three distinct but complementary projects provided valuable experience in rapid prototyping, API integration, and creating user-centric solutions.

The challenge of building practical tools for a specific industry domain was both challenging and rewarding, requiring deep understanding of user needs and technical implementation across multiple platforms.

Looking Forward

While these projects were built specifically for the hackathon, they demonstrate the potential for innovative solutions in the fine spirits industry. The combination of AI, computer vision, and browser automation shows how technology can enhance traditional industries and create better user experiences.

Overall, I'm incredibly proud of what I accomplished during the BAXATHON and the technical skills I was able to showcase across these diverse projects.