Projects

Owned Schema App's SOC 2 Type II compliance program end to end. Gap analysis, remediation across access controls, logging, and incident response, and managed the external audit. Also led KY3P compliance for major US banking clients.

Led the team that shipped Schema App's MCP Server, a secure gateway that exposes enterprise Content Knowledge Graphs to AI assistants including ChatGPT, Copilot, Claude, and Gemini. Designed the auth layer, permissioned entity access, and rate limiting. One of the first production MCP integrations in the structured data space.

Architected and deployed a fleet of AI-powered BDR agents that autonomously research prospects, personalize outreach, and manage multi-step email sequences. Integrated intent signals, CRM enrichment, and LLM-generated messaging to build pipeline at scale without growing headcount.

Built an agentic workflow system that acts as an AI chief of staff for engineering leadership. Triages Slack threads, summarizes PRs and incidents, drafts weekly stakeholder updates, and flags blockers across teams. Reduced exec overhead on status-gathering by 60%.

Led the team that built a RAG retrieval pipeline over the knowledge graph using LangChain, vector embeddings, and AWS Bedrock. Directed chunking strategies and reranking logic across 50K+ entities, with Langfuse for LLM evaluation, reducing hallucination in customer-facing AI integrations.

Owned the migration of TNGOC's core appraisal engine from on-premise to AWS. A 14-month initiative that moved 6 core services to cloud, cut annual operating costs by $1.2M, and improved platform uptime from 99.2% to 99.85%.

Directed engineering for Schema App's Entity Hub, a product that lets enterprise marketing teams identify, manage, and optimize entities across their content at scale. Integrated Google Search Console data with entity-level performance analytics so teams could measure AI search visibility per entity.

Oversaw the integration of NLWeb endpoints into Schema App's platform, enabling enterprise customers to expose their Content Knowledge Graph to agentic search experiences. Built the connection between structured Schema Markup and natural language interfaces, positioning the product for the shift from traditional search to agent-driven discovery.

Helped build Techcom's DRS SiVA, an omni-channel virtual agent for Canadian financial institutions. Handles customer identification, payment processing, inquiry resolution, and next-best-action recommendations in multiple languages, 24/7. Launched to production across banking and credit union clients.

Helped build Techcom's predictive next-best-action engine, combining reinforcement learning, causal AI, and customer lifecycle data to recommend the most effective action at each stage of the lending journey. Integrated with AI-prioritized worklists to operationalize model outputs across collections, retention, and cross-sell.

Drove a 19.7% AI Overview uplift for a flagship customer by restructuring entity graphs and aligning markup with LLM-parseable patterns. Published as a case study and became a key part of Schema App's sales motion.

Led the engineering effort to ship server-side Schema Markup rendering at the edge through Akamai and Cloudflare CDN integrations. Enabled enterprise customers to serve structured data without JavaScript execution, critical for AI crawlers and governance-heavy environments.