Senior AI Engineer who founded and leads Cotality's agentic AI platform — from the company's first LangChain backends to a production multi-agent ecosystem now used by 10+ engineering teams. Specializes in multi-agent orchestration, enterprise LLM systems, and AI platform engineering at scale. Deep GCP stack (GKE, VertexAI, ADK, Spanner); active U.S. security clearance through 2028.
Experience
Technical and engineering lead for Cotality's greenfield agentic AI platform, working alongside the enterprise architect to define AI strategy and deliver production systems. Founded the company's entire generative AI capability: built the first LangChain backends, led the first LlamaIndex agent framework, and drove the transition to Google ADK — platform now adopted by 10+ engineering teams.
- Built Cotality's generative AI capability from zero: first LangChain backends → first LlamaIndex agent framework → production Google ADK platform now used by 10+ teams across the organization.
- Designed and shipped an 8-agent production ecosystem integrated with 10+ MCP servers, with an intent discovery and routing architecture that sequences agents dynamically (e.g., data retrieval → analysis → visualization → output packaging) based on natural language input.
- Established organization-wide orchestration standards — hierarchical delegation, tool routing, inter-agent handoff protocols — adopted as the enterprise standard for agentic systems across Cotality.
- Personally built 3 of 8 production agents; mentored and guided the engineers who built the remaining 5.
- Led natural language-to-data product on GKE achieving 80% query accuracy on complex, multi-hop real estate requests using Spanner Graph DB and Redis Semantic Caching.
- Engineering lead for an enterprise Evaluation SDK built on Google ADK eval — integrates into Jenkins CI/CD pipelines, auto-generates evaluation reports, and exposes structured results to compliance, surveillance, and product consumers.
- Designing governance-as-code framework using OPA (Open Policy Agent) and Google ADK to enforce deterministic compliance guardrails throughout the agent lifecycle.
- Built Knowledge Store agent using VertexAI RAG Engine with vector search; implemented multimedia agent outputs including dynamic charts, maps, and Google Street View integrations.
- Presented AI-native engineering platform strategy to executive committee; nominated for Cotality's Pinnacle Award.
Built the MLOps foundation for Cotality's ML organization — standardizing how models are built, deployed, and operated across 6 business units on GCP. Inherited a fragmented landscape (local environments, inconsistent tooling, no shared standards) and delivered a production-grade, reusable pipeline framework.
- Designed Cotality's KubeFlow Pipelines (KFP) layer on VertexAI, creating reusable component templates for Apache Beam/Dataflow jobs, data preparation, model training, batch prediction, PubSub integration, model registry operations, and QC validation — reducing model deployment time by 30%.
- Supported MLOps infrastructure for 6 business units across models including AVM (property valuation), Home Price Index forecasting, water resiliency scoring, foundation type detection, and propensity models for HELOC and refinance.
- Transitioned legacy models to cloud-native GCP architectures, achieving 40% improvement in processing speed and 50% reduction in deployment cycle time.
- Engineered company-wide ML framework increasing engineering productivity by 25%; largest pipeline processes nationwide property data (hundreds of millions of records) monthly.
- Mentored junior engineers and authored onboarding documentation reducing new hire ramp time by 40%.
Most senior engineer on multiple concurrent defense client engagements. Built distributed data platforms and visualization frameworks for DoD clients dealing with fragmented legacy systems and high manual analysis burden. Holds active U.S. security clearance through 2028.
- Solved a critical data consolidation problem for a DoD client: built ingestion pipelines and scrapers to unify data from disparate legacy defense systems, eliminating 33% of manual data entry and replacing error-prone manual processes.
- Authored a reusable Angular visualization framework for consolidated defense data; demand spread organically to other divisions and organizations, contributing to $1M+ in follow-on contracts and reducing new project spin-up time by 50%.
- Architected secure distributed software systems increasing operational efficiency by 30%; engineered real-time Tableau cybersecurity dashboards for threat detection and situational awareness.
- Served as most senior engineer on multiple simultaneous engagements — responsible for architecture, code reviews, and quality of other engineers' work.
Sole data scientist at one of the Navy's primary research and test facilities, applying ML and scientific computing to sensor data analysis, performance modeling, and resource optimization for weapons systems R&D programs.
- Designed decision support systems using Python (Pandas, Scikit-Learn) to help R&D teams analyze sensor performance data and make resource optimization decisions for weapons systems development programs.
- Optimized MATLAB simulation reporting pipelines for sensor analysis, improving throughput by 10%; built automated reporting workflows replacing manual analysis processes.
Education
Awards
Nominated for Cotality's highest internal recognition for contributions to the company's agentic AI platform and AI-native engineering capabilities.
Received multiple spot awards recognizing high-impact contributions to AI platform delivery.
Invited to present Cotality's AI-native engineering vision and agentic platform strategy directly to the executive committee.