AI & data engineer
I build AI and data systems that hold up in production, and the full-stack products that put them to work.
I work across the model, API, data, and interface layers needed to turn an AI capability into a product a team can operate.
Today that means AI integrations and data solutions on enterprise cloud platforms: MCP servers, Copilot Studio agents, and Microsoft Fabric pipelines. Before that, agentic workflows and conversational AI. On the side, I build and run full products, including one that has recorded thousands of real business transactions.
Define the task, decision points, data boundaries, and failure states before choosing a model or agent pattern.
Test response quality, latency, recovery paths, and the cases where the system should hand control back to a person.
Authentication, permissions, data contracts, observability, and interface feedback matter as much as the model call.
Building MCP servers that connect enterprise assistants to Copilot Studio agents behind Entra ID, plus Azure AD SSO for Dynamics 365, Power Automate connectors, and FastAPI services on Microsoft Fabric.
Shipped multi-agent workflows, low-latency voice agents, and evaluated RAG systems, owning delivery from SRS through deployment.
Built Python data pipelines and a Next.js operations dashboard for monitoring large-scale e-commerce jobs.
Developed Angular applications, API integrations, and React-based SharePoint web parts.
A multimodal retinal-disease research prototype combining Fundus photography and OCT radiomics.
A production platform that runs a business's sales, purchasing, stock, and accounts, with its own WhatsApp Business app and AI-powered voice and chat entry.
An AI decision-brief generator for customer-care operations that ranks which contact drivers are worth automating.
A grocery price-comparison platform that scrapes major grocery stores, matches the same product across them, and answers shopping questions in plain English.
A QA workspace for teams that need test execution and defect follow-up to stay traceable.
Maps a candidate's skills to relevant roles, missing skills, and feasible bridge paths.
5 active · each one verifiable on Microsoft Learn
I can help with AI development, data engineering, analytics, and cloud, and the full-stack product around them.