tom kimingi

I’m a Systems Architect and Full-Stack Engineer with a strong background in SQL development and data architecture. I specialize in analyzing, editing, and writing complex SQL queries to streamline data workflows and support AI training tasks.

I build systems that make model pipelines more trustworthy and easier to debug. I’ve evaluated model performance and refined pipelines to enhance data integrity, with a focus on advancing AI models in cloud-based environments.

In my local-first RAG & knowledge graph project, I designed a fully local AI system using Incremental Knowledge Representation (IKR) and SQL/NoSQL best practices for efficient storage and retrieval. I implemented Triple Extraction (Subject-Predicate-Object) to transform unstructured documents into a queryable Neo4j graph, including truth-gated reasoning, and built a secure ingestion pipeline with SQL-based validation steps to prevent cloud data leakage.

I also create observability and automation into the engineering workflow. My AI Model Observability Tool uses OpenTelemetry and Model Context Protocol (MCP) to visualize real-time data flow and enable direct bug-querying and logic-tracing from within the IDE—reducing Mean Time to Detection (MTTD) for logic-based bugs by 60% through automated trace correlation.