I solve complex problems—sometimes with financial engineering, sometimes with software, often with both.

I won 2nd place with my good friend Valerie Yong at the Rosewood Sand Hill Hospitality 2030 Hackathon, organized by Cerebral Valley. The event, sponsored by Rosewood, Anthropic, ElevenLabs, and Greycroft, focused on how luxury hotels can apply AI to improve and refine guest experience across a global portfolio.

Our core product principles were:

1. Don't tax the customer with forced data intake (paperwork) or web scraping; always solve for creating an effortless guest experience.

2. Luxury hospitality is a global business supported by international teams. In a highly personal, people-first setting, AI can remove language barriers to improve staff choreography, enabling each team member to deliver the highest-quality experience with every interaction.

A sweet full-circle moment: I entered and won my first hackathon a little less than a year ago, and have placed in six total since — each an accelerated step on the path to becoming technical, working alongside experienced engineers to ship my own ideas fast.

Most agentic-commerce demos run in one direction: agents paying agents. Gyasss runs in both directions on the same Circle Nanopayments primitive — users earn USDC for confirming gas station prices, AI agents pay $0.001 per query for the data.

The build pushed me deep into Circle, Arc, x402, EIP-3009, and nanopayments for AI agents. The most interesting design problem was the trust layer — building an information marketplace where contributors are economically rewarded for providing useful, accurate data, with consensus weighting, time decay, stake-and-slash, and outlier detection protecting against grinding and Sybil attacks.

The economic primitive is the unlock. On Ethereum mainnet, gas alone runs 1000x the per-query price. On Stripe, the 30-cent minimum fee is 300x. Arc with Circle Gateway is what makes sub-cent settlement positive-margin on both sides.

I built Gyasss solo over one week as a small-scale test of an architectural idea I've been thinking about: that productized, verifiable economic primitives could replace the person-to-person information networks that gatekeep value in industries built on non-public data. Real estate is one such industry. This was the proof of concept.

MARA Holdings is the second-largest corporate holder of Bitcoin. They challenged participants to build an AI-driven system to arbitrage energy and AI inference prices across their global data center portfolio.

My team built Panoptic, a platform that integrates satellite data with real-time energy pricing to optimize compute allocation and infrastructure investment. The system includes a global visualization of MARA's data centers, AI-powered analysis for battery storage buildouts and expansion siting, an automated energy treasury that deploys power based on grid pricing, and a derivatives platform for hedging Bitcoin and energy exposure.

The derivatives component drew directly on my background in interest rate risk—the same hedging frameworks that work for real estate debt also work for energy and crypto treasury.

Traditional A/B testing is slow and manual. I designed a system that applies evolutionary biology to user interfaces—beneficial changes propagate, harmful ones die off, and the UX evolves autonomously.

Survival of the Feature uses two AI agents working in coordination. The first generates UX variations and deploys them to a sample of users. The second applies statistical analysis to real-time user response, determining whether each variation improves engagement while calibrating how quickly to expand exposure. It's similar to natural selection, except data accelerates the process.

We won first place out of 45 teams and presented at the conference keynote.

My first hackathon. Built an automated rooftop inspection system that analyzes aerial drone footage to detect property issues including cracks, missing tiles, leaks, and debris. The tool extracts frames from drone videos and processes them through Google Gemini's vision API to identify maintenance concerns at scale.

Technical approach:

• Python pipeline using ffmpeg for frame extraction at configurable intervals
• Batch processing system for analyzing hundreds of images from a single flight
• JSON response consolidation for downstream reporting and visualization
• Bash orchestration for production deployment

Impact: Won first place at the Aparavi AI Hackathon, demonstrating practical applications of computer vision for real estate due diligence. The tool addresses a significant pain point in property assessment—manual roof inspections are time-consuming, expensive, and often miss issues visible only from above.

Stack: Python, Google Gemini API, ffmpeg, bash

Construction site monitoring typically costs $1,500 per visit — about $18,000 annually per project — and the information is delayed and incomplete by the time it reaches stakeholders. You're paying for monthly snapshots when you need a continuous feed.

SpotCheck is a multimodal AI agent that analyzes daily drone footage to monitor construction sites in real time. The system generates 3D site models, runs automated OSHA compliance scanning between scheduled inspections, and flags delays and potential cost overruns early enough to intervene. Users can speak with the agent conversationally — querying specific areas of a site, comparing progress against timeline, or pulling up safety concerns — instead of waiting for a consultant's next visit.

The hackathon required chaining at least three sponsor tools via MCP (Model Context Protocol) into an end-to-end agentic workflow. We built SpotCheck on AWS with Minimax for video understanding, BEM for 3D reconstruction, and Operant for agent security — connecting drone capture to spatial analysis to a conversational interface a project manager could actually use.

This was the second drone-powered real estate project in two weeks, following our first-place win at the Aparavi Hackathon with Roofi. The through-line: aerial imagery and computer vision applied to inspection problems that are currently manual, expensive, and infrequent.