About
Bio
I am an external consultant and senior technical advisor who specializes in designing intelligent, scalable platforms that elevate customer experience and operational efficiency. My consulting practice leverages deep expertise in architecting advanced AI systems, with a focus on large-scale search, semantic vector databases, cognitive ranking, fine-tuning LLMs, creating classifiers, training embedding models, and retrieval-augmented generation (RAG). I am proficient in GPU-accelerated model training and fine-tuning on NVIDIA DGX Spark, including SFT/GRPO workflows with PyTorch and Unsloth for domain-adapted LLMs and vision models.
Clients engage me to align AI-driven innovation with enterprise goals, optimize performance, and enhance knowledge retrieval while ensuring sustainable technology adoption. I also provide leadership coaching and technical mentorship for cross-functional teams. Based in Austin, Texas.
Experience
Independent Consulting — AI Architecture, Agentic Workflows, ML Solutions
Self-Employed · May 2025 – Present · Austin, Texas
Designing and implementing agentic workflows, MCP-based solutions, RAG systems, and LLM fine-tuning for a range of clients. See Clients section below.
CTO and Distinguished Engineer
IBM · January 2018 – May 2025 · Tampa/St. Petersburg, Florida
- Architected large-scale Elasticsearch solutions for IBM client-facing sites
- Designed vector databases with Watsonx.data and Milvus for semantic search
- Built scalable vector stores with PostgreSQL (pg_vector, pg_vectorscale)
- Integrated LangChain for RAG and agent workflows
- Combined watsonx.ai and LLMs with search for chatbot and content retrieval
Chief Globalization Architect
IBM · December 2013 – January 2018
Led IBM's globalization strategy for cloud and mobile products. Architected the Globalization Pipeline service on IBM Bluemix. Developed global cloud solutions using Watson cognitive services.
Additional roles at IBM: Senior Technical Staff Member, Lab Assignee, Advisory Software Engineer, Staff Software Engineer. 19+ years at IBM total. Ph.D. in Computer Science, Florida Institute of Technology.
Selected Clients
| Client | Engagement |
|---|---|
| SiteZeus | Agentic workflows leveraging MCP servers |
| Decked Out Factory | Customer support chat with n8n, MCP, OpenAI, Anthropic; vector data stores for semantic + lexical search |
| RexCare | Customer support chat with agentic workflows |
| PSA Staffing | Automated time tracking, AI-driven insights and forecasting |
| Cultara | Semantic search; SFT/GRPO fine-tuning with Unsloth + PyTorch (DGX Spark) |
| ZW Consulting | Domain-specific vision models with Unsloth (DGX Spark) |
| Commercial Acoustics | Prompt workflows for proposal analysis and bid/no-bid decisions |
Training & Certifications
25+ certifications from DeepLearning.AI, Hugging Face, NVIDIA, LangChain, Coursera, IBM, and Udemy. Representative areas:
- LLMs & Fine-tuning: Pretraining LLMs, Fine-tuning & RL for LLMs, Post-training of LLMs, Mathematics Behind LLMs and Transformers, Fine Tune OpenAI GPT Models
- PyTorch: PyTorch for Deep Learning, PyTorch Fundamentals, PyTorch: Advanced Architectures and Deployment, PyTorch: Techniques and Ecosystem Tools
- Quantization & Optimization: Quantization in Depth, Quantization Fundamentals (Hugging Face), Quantization for GenAI Models
- Agents & MCP: AI Agents Fundamentals (Hugging Face), Fundamentals of MCP, Model Context Protocol Masterclass, Serverless Agentic Workflows (Amazon Bedrock)
- Search & RAG: Embedding Models: from Architecture to Implementation, Semantic Caching for AI Agents
- Safety & Reliability: Red Teaming LLM Applications, Safe and Reliable AI via Guardrails
- Other: Generative AI with Diffusion Models (NVIDIA), Amazon Bedrock Complete Guide, Introduction to LangGraph (LangChain), n8n AI Agents, IBM Design Thinking Practitioner
AI Experience
Deep hands-on experience across the AI stack:
- Search & Retrieval: Elasticsearch, vector databases (Milvus, pg_vector, pg_vectorscale), ParadeDB, hybrid semantic + lexical search
- RAG: Retrieval-augmented generation, LangChain integration, Watsonx.data
- LLM Fine-tuning: SFT, GRPO, PPO; Unsloth + PyTorch on NVIDIA DGX Spark
- Vision & Generative: Domain-specific vision models, diffusion models
- Agentic Workflows: MCP servers, n8n, OpenAI, Anthropic, Amazon Bedrock
- Tools: LangGraph, Hugging Face Transformers, DuckDB, OpenSearch
Master Inventor at IBM (30+ patent applications). Honorable mention for best dissertation; First Place Paper Award from Sigma Xi. Publications on cloud-native services, globalization, and Unicode.