AI Agent Developer & Fintech Automation Builder Spirano, Italy · Building autonomous systems at the intersection of LLMs, blockchain and financial infrastructure.
I design and ship multi-agent architectures, on-chain automation, and AI-driven trading tooling. My work focuses on systems that act, not just analyse — agents that hold state, coordinate, and produce measurable outcomes.
- Email: yeisonriascos@gmail.com
- LinkedIn: linkedin.com/in/yeison-riascos-sanchez
- GitHub: github.com/yeick010
- ArbitrageAgent — AI-powered cryptocurrency arbitrage system with paper-trading dashboard (private).
- AgentShield-MCP — multi-agent orchestration platform with validator contracts (private).
- Portfolio Orchestrator — unified ecosystem for coordinating multiple AI agents (private).
- Croméyeuï Cafè Bot — WhatsApp business automation, Mistral API integration (private).
- Proof-of-Coherence — public research artefact on falsifiability frameworks for AI claims — repo.
Languages: Python, JavaScript / Node.js, Bash, Solidity AI / Agents: LLM integration, prompt engineering, CrewAI, LangChain, multi-agent coordination Cloud: Google Cloud Platform (Cloud Run, Cloud Shell, Secret Manager) Web3: Coinbase CDP SDK, Thirdweb, Alchemy, Circle, chain-agnostic design Data & BI (foundational): Python (Pandas, NumPy), SQL, Power BI, NLP
- First-principles. Understand the constraint before optimising the solution.
- Pre-registration over post-hoc. When I run an experiment, the falsifiability conditions are written before the data is seen.
- Cost-aware. I optimise for the cheapest path to a real measurement, not the most elegant one.
- Honest about scope. I'd rather ship a small thing that works than promise a large thing that doesn't.
- Collaboration on multi-agent systems and autonomous trading infrastructure.
- Conversations with teams working on agent observability, validator design, or AI-finance integration.
- Research-adjacent roles where engineering rigour matters.

