Yoree Signals Market

• 

  Meet Yoree, your trading sidekick. :)

Inspiration

Traders trade, agnostic of asset classes. This MO is something that I have seen percolate within me as well as an emerging trend amongst serious traders. It was this inspiration that motivated me to create a strategy gen platform in the Web3 space and further expand my project into a full signals market. Before I created YSM, I spent hours manually searching exchanges, analyzing correlations, and piecing together opportunities from fragmented sources. This labor of love was not in vain. I utilized my dev skills and my intimacy with the pre-markets to build out the vision I had for what I see as the true future of trading: real-time, cross-asset class trading wrapped in a singular platform. The problem isn't a lack of data; it is the reimagining of a generalizable XAC (cross-asset class) trader's desk: start with a signal hypothesis and let the system surface related assets automatically.

What it does

Yoree Signals Market is a signal-driven intelligence exchange where traders create signals (e.g., "Venezuela Oil" or "ETH adoption") and the platform automatically discovers and connects to related assets across crypto, stocks, futures, forex, ETFs, bonds, and commodities; all in one unified interface. The platform uses a 3-agent pipeline (Data Feed Selector, Hypothesis Generator, and Asset Discovery Agent) to transform raw input into tradeable signals with automatic cross-asset connections, complete with direct exchange links for immediate execution. Users can trade signal quality as first-class assets, with continuous scoring that updates based on performance, accuracy, and market consensus.

How I built it

I built Yoree as a modular system with three layers: Minam (data normalization and feed management), Syuzhet (narrative generation and hypothesis structuring), and the Yoree Signals Market (orchestration and asset discovery). The frontend is React/TypeScript with Material-UI; the backend is Rust (Axum) for performance. I integrated Google's Gemma for AI-powered asset discovery and strategy generation, and built a signal scoring system that tracks quality over time. The platform orchestrates these components through a RESTful API architecture, with WebSocket support planned for real-time updates. Smart contracts on BNB Chain and Arc Testnet handle signal market mechanics, while the AI agents work together to transform unstructured input into structured, tradeable signals.

Challenges I ran into

The biggest challenge was creating a reliable asset discovery system that works across asset classes with different data cadences; crypto moves in minutes while commodities trend over weeks. I had to design a system that could normalize these temporal differences while maintaining accuracy. Another challenge was integrating multiple blockchain networks (BNB Chain, Arc Testnet) while keeping the user experience seamless. Finally, building the AI orchestration layer required careful prompt engineering to ensure the agents discover relevant assets and generate actionable strategies, not just generic recommendations. The solution was creating a specialized model architecture that layers asset-class-specific data streams and dynamically adjusts information processing cadence based on real-time market behavior.

Accomplishments that I'm proud of

I'm proud of winning the 2025 Blockworks Permissionless IV Hackathon and being a featured finalist at the NYC AI Tinkerers Agentic AI App Hackathon with Google Cloud Run GPUs. I've successfully integrated two production-ready open-source components (Minam and Syuzhet) into a cohesive platform, demonstrating the ability to build complex, integrated systems that combine AI agents, real-time data feeds, and blockchain technology. Most importantly, I've created a working system that inverts the traditional trading workflow, proving that signal-first asset discovery is not just possible, but more efficient than the status quo. The platform is preparing for launch on BNB Chain in Q1 2026. But the real-world use case is what I am most proud of...I use Yoree Signals Market for my own personal trading.

What I learned

Building Yoree taught me how to integrate AI agents, real-time data pipelines, and blockchain infrastructure. My work as an AI Agent Engineer taught me how to develop backtesting modules and deploying SDKs but YSM is a whole different level of complexity. I learned to design systems where specialized components (Minam for data, Syuzhet for narratives) work together seamlessly. I also learned that asset discovery requires understanding correlations across asset classes: crypto, stocks, futures, forex, ETFs, bonds, and commodities; each with different cadences and behaviors. This led to designing a malleable ML framework that adapts temporal processing to how markets actually move, rather than forcing a one-size-fits-all approach. I gained deep experience in prompt engineering for AI orchestration, understanding how to balance specificity and flexibility to generate actionable outputs.

What's next for Yoree Signals Market

In the next 6 weeks, I'm prioritizing real-time WebSocket integration for live signal updates, advanced ML-based feed selection for AI agents, and enhanced asset discovery algorithms to improve correlation analysis across asset classes. I plan to build a specialized model that layers asset-class-specific data streams to dynamically adjust the cadence of general market information, creating a dynamic training framework that adapts to how different markets actually move. I'm also developing a personalized recommender system that improves time-on-the-app metrics by channeling users to trading theses aligned with their preferences, risk tolerance, and historical trading patterns, ensuring each user discovers signals most relevant to their strategy. Long-term, I'm focused on launching on BNB Chain and Circle's Arc Network expanding to institutional features, cross-chain strategy deployment, and building a strategy marketplace with monetization capabilities. Eventually, I intend to build my own proprietary model that specializes in layering data based on asset classes, making the system more responsive to actual trading dynamics while the recommender system continuously learns from user behavior to surface increasingly relevant opportunities.

Built With

• ai-platform)-supabase-databases:-postgresql-(planned)-mock-database-(mvp)-apis-&-data-sources:-coingecko-api-google-trends-api-twitter/x-api-reddit-api-news-apis-(bloomberg
• axum
• coingecko
• docker
• gcp
• gemma
• googletrends
• javascript
• nextjs
• openzeppelin
• react
• reddit
• rust
• solidity
• sql
• typescript
• x