I build agentic systems that solve real-world problems reliably.

My foundation is a BSc in IT specializing in Data Science, which grounded me in software, data and how these technologies serve business. My MSc in AI followed. My dissertation, AgentC, introduced me to multi-agent systems (MAS) and how they tackle complex problems. AgentC reduces Scope 2 GHG emissions safely and without sacrificing comfort, using a distributed MAS that perceives and controls physical electrical devices. It is a hierarchical MAS in LangGraph: a manager agent coordinates nine task workers and a planning sub-graph of four expert agents, with shared Pydantic state, checkpointing and LangSmith tracing.

Professionally, I've spent two-plus years building enterprise software for data-heavy production needs, such as emissions reporting and GIS visualizations. Notably, I designed and shipped a visual AI quotation service for electrical and solar work. It uses multimodal models to return structured assessments, delivered over a messaging service. That project taught me that the observability and explainability of decisions matter more than raw model output. Production work across DRF, FastAPI, PostgreSQL, React, Next.js, third-party integrations, Stripe, AWS and CI/CD gave me the architectural toolkit to design production agentic AI systems.

Right now, I'm building a fabrication quotation agent for a real use case of my own. My goal is to join teams building impactful intelligent systems that keep humans and machines working together.