Company-wide GenAI technical lead at Bio-Techne — architecting and shipping multi-agent systems, DSPy-optimized RAG pipelines, and evaluation frameworks used by 300+ people weekly. Databricks AI Days speaker. Building AI that actually gets deployed.
I architect, ship, and own GenAI systems in production — not demos, not POCs, but infrastructure that hundreds of people rely on daily.
As company-wide GenAI technical lead at Bio-Techne, I own the full stack: architecture, deployment, governance, and operations of AI systems used by 300+ people every week. From multi-agent orchestration to LLM evaluation frameworks that catch hallucinations before they ship.
My approach? Start small, iterate fast, ship to production. The engineering patterns — multi-agent orchestration, RAG at scale, LLM evaluation — apply across any industry. The life sciences context is where I've sharpened them, not where they're limited.
Company-wide GenAI technical lead. Shipped 8 production AI systems across multi-agent, RAG, and ML. 300+ weekly active users, 85% support deflection rate. Databricks AI Days speaker (300 attendees). Mentored 3 GenAI engineering interns.
Built and deployed a graph-based product recommendation system (20% CTR in A/B testing) and a text classification model that reduced misrouted Salesforce cases by 90%.
Dual master's degrees in AI and Data Science. First-author publication in Springer CCIS. Keen Scholarship recipient.
Invited to speak on building and deploying multi-agent GenAI in life sciences — taking agentic systems from prototype to production on Databricks.
I walked through how my team at Bio-Techne built and shipped an orchestrated multi-agent system — combining web research, transaction analysis, and marketing intelligence on Databricks to surface untapped product opportunities for the sales organization.
A curated selection of AI systems I've architected and deployed
Sales Intelligence Platform
An orchestrated multi-agent system combining web research, transaction analysis, and marketing intelligence to identify untapped product opportunities. Built with LangGraph for complex agent coordination.
Production RAG for Support Teams
A production multi-agent retrieval system searching Salesforce cases and an Egnyte knowledge base. Migrated to DSPy with MIPROv2 prompt optimization and CRAG for corrective retrieval. LLM-as-judge evaluates both retrieval accuracy and answer quality before responses ship.
Customer Service Response Engine
An agentic system that drafts answers to lengthy customer questionnaires by retrieving from product documentation and prior responses — turning a manual, multi-day task into a same-day workflow for the customer service team.
Million-Scale PDF Processing
Template-based extraction system processing ~1M Certificate of Analysis PDFs with high accuracy. Automated data validation and structured output for downstream analytics.
Semi-Supervised Learning Pipeline
Teacher-student learning model with XGBoost to predict optimal capture-detect antibody pairs. Leveraging SPR data from Carterra and Ella immunoassay platforms for training.
Internal AI Platform with LibreChat
Deployed LibreChat as an internal ChatGPT alternative, providing secure AI assistance across the organization with custom integrations and enterprise authentication.
Customer Care Intelligence Platform
Ingests customer questionnaires, extracts each question, retrieves answers from the vector index, and highlights the exact source passage using PDF coordinates — building trust by showing provenance for every answer. When the knowledge base has gaps, users upload new sources that trigger an automated ingestion and embedding pipeline, making the system self-healing.
Knowledge-Graph-Powered Wiki Search
Built a graph-based retrieval layer for Bio-Techne's internal wiki, where pages cross-reference each other. A flat vector index would miss those connections and be prohibitively expensive to maintain. The graph structure traverses references at query time, surfacing multi-hop answers that dense retrieval alone cannot reach — now powering one of the Technical Service Agent's core retrievers.
Network-Based Product Discovery
Graph-based recommendation pipeline leveraging product relationships and customer behavior patterns to surface relevant suggestions with explainable reasoning.
Hover or tap to explore my technology stack
Have questions? Want to schedule a call? Chat with my AI — it knows my schedule and can help you connect.
Whether you have a project in mind, want to discuss AI opportunities, or just want to chat about the future of biotechnology or AI in general — I'd love to hear from you.