Itinitek

William Foland, PhD

AI Executive & Architect

I build production AI platforms, and the teams and strategy around them. Healthcare AI, agentic systems, and clinical workflow automation. Former VP, Cirrus Logic. PhD in AI. 25 patents.

Get in Touch See the Work

About

I'm an AI executive and architect who builds production AI platforms and the teams and strategy around them.

My first career was in semiconductor digital design. I was one of a small group of technical VPs at Cirrus Logic, a 4,000-person company, and later Senior Director at Marvell, leading an international team of 150 engineers. I brought numerous products to market and earned 25 patents along the way.

I then earned a PhD in AI at CU Boulder, focused on deep learning for extracting meaning from language, with research presented at the Association for Computational Linguistics. Today, I combine that AI expertise with the executive discipline of decades spent shipping complex products—building systems for clinical applications, Harvard Business School, and industrial engineering.

Engines

Production systems and the architecture behind them. Each card is a look at the real work — clinical AI platforms, real-time orchestration, and vision pipelines engineered to run reliably under load.

Clinical Evidence Platform · in production

Clinical documents → structured, computable evidence

A production AI platform for a leading clinical genetics laboratory, architected and delivered from scratch in 12 weeks on a HIPAA-compliant cloud. It turns dense, unstructured records into evidence that genetic counselors act on with confidence, supporting a significant reduction in case turnaround time.

  • Automated extraction. Turns lengthy, unstructured records into organized clinical findings.
  • Evidence traceability. Links findings to source pages and highlighted passages for verification.
  • Expert oversight. Clinical reviewers validate, correct, and approve results.
  • Integrated review. Summaries, findings, terminology selection, and source documents in one workspace.
  • Structured output. Standardized, reviewed information for downstream clinical workflows.
  • Built around clinical practice. Developed through close collaboration with clinical experts.
  • Clinical Genomics
  • ICD-10
  • HIPAA
  • Human-in-the-loop
Spatial-Semantic Pipeline

2D ISO drawings → 3D system model

A vision-LLM pipeline that ingests a stack of flat 2D isometric drawings and reconstructs them into a single validated 3D system model. Perception (vision/OCR) is separated from reasoning (LLM) so each stage tunes independently, with geometric consistency checks driving error rates to safety-critical levels — then you chat with your system in plain English: "find all butterfly valves in the 2-inch line." Weeks → hours.

  • Vision-LLM
  • OCR
  • Geometry Checks
  • Oil & Gas
Real-Time Multimodal Orchestration

One reasoning model driving four live avatars

A real-time meeting simulator where a single LLM controlled four distinct video avatars at once — coordinating speech, timing, and behavior under tight latency over WebRTC + Next.js + FastAPI. Speech-driven scenarios allow active student participation in the conversation. Delivered to 900+ MBA students; a closed-loop eval system stress-tested coordination with synthetic dialogue.

  • WebRTC
  • Multi-Agent
  • Low Latency
  • Harvard
Governed Agents · Control by Design

The right intelligence for each decision. Controls on every action.

Reliable AI requires more than a capable model. This architecture combines explicit policies, focused decision models, language reasoning, and human review—matching each task to the appropriate method and checking proposed actions before execution.

  • Deterministic checks. Enforce permissions, scope, and prerequisites in code.
  • Selective reasoning. Use models where interpretation is needed and language models where language is the task.
  • Human escalation. Route unresolved or consequential decisions for review.
  • Traceable actions. Record the evidence, checks, and decisions behind execution.
  • Policy as Code
  • Decision Models
  • Human Review
  • Audit Trail

Engagement Models

Four ways to work together — from running your AI program to dropping a proven engine into your cloud.

AI Leadership

Fractional or full-time

Own the AI roadmap, build and lead the team, set governance for regulated data (HIPAA), and get platforms into production. Executive discipline from decades of shipping hardware, applied to frontier AI.

Discuss leadership

Strategic Advisory

Retainer-based

Architecture roadmapping, multi-model cost/latency routing analysis, vendor vetting, evaluation strategy, and security governance for agent action spaces. Async-first strategic oversight plus scheduled executive syncs.

Discuss advisory

Embedded Architect

Project-based

Direct ownership of engineering teams — designing planner-executor-verifier self-correcting loops and deterministic code-as-judge constraint checks. Active, project-bound engineering leadership from prototype to production.

Discuss a project

Deployment

Accelerator integration

Licensing, optimization, and deployment of a proven engine — the Clinical Evidence or Spatial-Semantic core — directly into your cloud environment. A focused 90-day execution sprint that bypasses standard R&D timelines.

Discuss deployment

Core Expertise

Agentic AI Systems

Multi-agent orchestration for production workflows, including planner-executor-verifier loops, MCP tool use, and stateful versus stateless reasoning strategies.

  • Multi-Agent
  • MCP
  • Verification Loops
  • Retry Logic

Evaluation & Reliability

Closed-loop evaluation systems, failure-mode analysis, and iterative tuning for AI products that need to work reliably under real constraints.

  • Dataset Construction
  • LLM-as-Judge
  • Guardrails
  • System Tuning

Generative AI & LLMs

RAG pipelines, fine-tuning, prompt engineering, and model routing for domain-specific systems that improve expert workflows instead of adding brittle automation.

  • RAG
  • MCP Servers
  • PEFT / LoRA
  • Prompt Engineering

Real-Time Multimodal

Low-latency voice, video, and avatar systems built for human interaction, teaching simulations, and other applications where timing and responsiveness matter.

  • WebRTC
  • Voice / Speech
  • Video Avatars
  • Low Latency

Computer Vision

Vision pipelines for industrial and edge workloads, combining detection, OCR, parsing, and multimodal reasoning to recover usable structure from complex visual inputs.

  • YOLO
  • OCR
  • Video Analytics
  • Multimodal Reasoning

Infrastructure & Delivery

End-to-end system design spanning model serving, cloud infrastructure, and application delivery across prototypes, pilots, and production systems.

  • AWS / Azure
  • PyTorch
  • TensorFlow
  • FastAPI / NextJS
  • Python / C++ / JS

Credentials

PhD Neural Language Understanding
25 US Patents
20+ Years Production Delivery

Education

PhD in Computer Science

University of Colorado, Boulder, CO (December 2017)

Dissertation: Natural Language Understanding: Deep Learning for Abstract Meaning Representation

Deep learning models similar to those that power ChatGPT and modern LLMs.

MS Computer Science

University of Colorado, Boulder

BS Electrical Engineering

University of Colorado, Boulder

Selected Publications

Full history & patents Two careers across silicon and AI

Before pioneering GenAI solutions, I spent years architecting mixed-signal integrated circuits and read channels for data storage. That foundation in low-level signal processing, error correction, and hardware constraints directly informs how I build efficient, scalable, robust AI systems today.

May 2026 - Present

AI Architect (contract)

Leading clinical genetics laboratory, remote

  • Architected and delivered, from scratch in 12 weeks, a production AI platform that converts clinical documents into structured evidence for genetic counselors, supporting the lab's push to significantly reduce case turnaround time.
  • Led an AI opportunity assessment across lab operations, shaping the organization's AI roadmap.
  • Built on a HIPAA-compliant cloud environment, working within the lab's governance and access controls.
Jan 2026 - Mar 2026

Founder & Chief Consultant

Itinitek Ltd, Golden, CO

  • Designed an ingestion engine combining specialized AI agents with deterministic engineering checks so extracted data is verified against physical and geometric constraints before landing in the database.
  • Built workflows to route ambiguous cases for fast human review while preserving provenance and validation history on every extracted value.
Mar 2023 - Sept 2023

Founder & Chief Consultant

Itinitek Ltd, Golden, CO

  • Fine-tuned LLaMA for the legal domain in 3 weeks and built a RAG pipeline that handled 500+ page contract sets.
  • Built an AWS pipeline combining LLMs and SQL for real-time options analysis; the proof of concept helped drive a client funding round.
July 2020 - Feb 2023

Chief Scientist

Lilac Cloud Inc, Cupertino, CA

  • Architected vision-AI pipelines for edge computing, combining Dockerized GPU services with CUDA-accelerated model execution for real-time video analytics in constrained environments.
  • Built custom FFmpeg filters and Libav extensions to run inference inside the video pipeline, enabling object recognition and event detection without interrupting frame flow.
  • Implemented GAN-based imperceptible video watermarking and delivered low-latency frame processing for live sporting events.
May 2018 - June 2020

Cofounder & Chief Scientist

Bolt Analytics Corp, Santa Clara, CA

  • Architected time-series anomaly detection for financial services using CNN, RNN, and transformer approaches.
  • Led an offshore team building automated diabetic retinopathy detection from retina scans using TensorFlow.
Aug 2016 - Apr 2018

Research Scientist

CU Computational Language and Education Research, Boulder, CO

  • Developed recurrent NLP models for automated speech recognition and dialog analysis in K-12 STEM education.
  • Built a system to help teachers reflect on and improve instructional practices through AI-powered feedback.
May 2012 - July 2014

Expert Technical Consultant

Dovel & Luner, LLP, Santa Monica, CA

  • Served as an expert witness in semiconductor patent litigation, contributing technical analysis that supported successful client outcomes.
May 2009 - May 2012

Founder & Independent Developer

Itinitek Ltd, Golden, CO

  • Architected, developed, and marketed six GPS, graphics, and skiing applications for iPhone (Objective-C), concurrent with MS and PhD coursework at the University of Colorado.
Nov 2003 - May 2009

Senior Director, Optical Products

Marvell Semiconductor Inc, Santa Clara, CA

  • Led a 150-engineer division developing full SoC solutions spanning ARM core, servo, read channel, and DSP on one chip.
  • Managed the complete product lifecycle from architecture through silicon to production across multiple generations of mixed-signal optical storage controllers.

U.S. Patents

8,559,287 Method and System for Fault Protection Using a Linear Feedback Shift Register
8,498,186 Circuits, Architectures, Apparatuses, Systems, Algorithms and Methods and Software for Optimum Power Calibration for Optical Disc Recording
8,391,115 Method of Improving Quality of Optical Recording Using Circumferentially Repeatable Compensation
8,345,523 Method and Apparatus for Optimizing Optical Recording
8,060,674 Systems and Methods for Data Storage Devices and Controllers
7,475,173 Integrated Disc Drive Controller
7,379,452 Synchronous Read Channel
6,594,716 Mixed-Signal Single-Chip Integrated System Electronics for Data Storage Devices

25 U.S. patents spanning signal processing, read channel design, mixed-signal integrated circuits, and optical recording systems. Full patent list available on request.

Let's architect your AI system.

Available for fractional GenAI leadership, architecture, and engine deployment.

bill.foland@itinitek.com (303) 601-8905 LinkedIn Profile