AI Agent Engineer

I deliver AI systems end to end, implementing LangGraph workflows, pgvector-based RAG, and Langfuse tracing to monitor latency, token costs, and response quality.

I built LedgerJob, a TypeScript AI business-operations agent using the OpenAI Responses API with multi-step tool execution, SSE streaming, persistent conversation history, and token-usage tracking. I exposed 60 business tools through an internal agent and shared MCP server, covering leads, quotes, invoices, projects, appointments, QuickBooks, messaging, and service agreements, with confirmation gates and bounded execution loops. I also built Cape Cod Permits, a scheduled permit-data collection platform for 15 municipalities that normalizes data into PostgreSQL and provides MCP tools for permit search, statistics, contractor profiles, coverage guidance, scoped contact access, and contact-query audit logs.

I built LangChain, Ollama, and n8n pipelines integrated with internal data sources and APIs. I implemented document-grounded RAG systems using nomic-embed-text embeddings and pgvector.

I operated high-availability Linux and database infrastructure. I also automated server provisioning and operational tasks using Ansible.

I supported critical healthcare infrastructure in an environment governed by strict service-level agreements.

My responsibilities included Active Directory administration and ServiceNow-based support operations.

I administered Linux and Windows server environments, supporting core infrastructure operations.

I also worked with firewalls, Hyper-V virtualization, and automation scripts to improve infrastructure management.

In progress.

Linux Fundamentals certification.

Technical Leadership certification.