William Klein
Resume

William Klein

Analog and mixed-signal engineer applying AI and machine learning to engineering problems.

View projects Download resume
A second-order step response, the waveform circuit designers check constantly. Change the damping and watch the overshoot.
Overshoot
31%
Settles within 2% at
 
Response
Underdamped

About

Analog and mixed-signal engineer with 18 years of experience, from analog design support to direct analog and AMS design, now applying AI and machine learning to engineering problems. During a career sabbatical living aboard a sailboat in the Caribbean, I kept my engineering skills sharp through hands-on work on complex marine electrical systems, power management, and system troubleshooting, while building proficiency in Python and AI/ML. I then completed UT Austin's Post Graduate Program in AI & Machine Learning while working in retail and as a scuba instructor, adding skills in predictive modeling, neural networks, computer vision, and generative AI. I bring seasoned analog expertise together with practical AI capabilities, including a circuit-simulation agent I built, to forward-thinking engineering teams.

William Klein

Projects

spice-agent

An AI agent that runs ngspice circuit simulations. Python and PySpice drive the simulator, and an MCP server exposes it to any MCP-compatible agent. I tested local Ollama models and chose Google Gemini to run the agent for reliable tool calling. Open source under the MIT license.

  • Python
  • ngspice
  • PySpice
  • MCP
  • Google Gemini
View spice-agent on GitHub

Coursework projects

From the UT Austin Post Graduate Program in AI & Machine Learning.

NLP_RAG

A retrieval-augmented generation assistant that answers clinical questions from a medical reference of about 4,100 pages.

  • RAG
  • LLMs
  • Python
View on GitHub

ReneWind

A neural-network classifier that flags wind turbine generators likely to fail, from 40 sensor readings. Tuned for recall because a missed failure costs more than a false alarm, reaching about 86.5% recall on the test set.

  • Neural networks
  • Predictive maintenance
  • Python
View on GitHub

SuperKart

Forecasts revenue for a product in a store for a retail chain, packaged as a Streamlit app and deployed to a Hugging Face Space.

  • Forecasting
  • Streamlit
  • Hugging Face
  • Docker
View on GitHub

  • EasyVisa

    Predicts whether a U.S. foreign-worker visa application will be certified or denied.

    View on GitHub

  • Personal_Loan_Campaign

    A decision-tree model that predicts which bank depositors will accept a personal loan offer, and which traits drive that choice.

    View on GitHub

Skills

Select a skill to see the projects, roles and coursework where I used it.

AI and machine learning

Analog design, test and EDA

Teams and customers

No skill selected.

Experience

  1. 2025 – 2026

    Retail Sales Associate & Scuba Instructor · Scubaland

    • Taught scuba certification courses and advised retail customers on dive equipment selection
    • Helped set up and customize a new CRM system for customer and sales tracking
  2. 2015 – 2017

    Analog Design Consultant (Independent Contractor) · Exar Corporation

    • Utilized Cadence Virtuoso for analog implementation of a DAC
    • Performed advanced circuit simulation using Spectre simulator
    • Delivered comprehensive design solutions using integrated analog design flow
  3. 2013 – 2015

    Staff Engineer, Project Lead · Exar Corporation

    • Led a cross-functional collaboration of 12 engineers for Mixed-Signal Design projects, including a Sensor Interface AFE, using Cadence AMS Designer
    • Collaborated with other teams to meet physical design challenges
    • Implemented rigorous DRC/LVS verification procedures and ran post RC extraction simulations for physical design sign-off
    • Mentored team members in advanced use of EDA tools
  4. 2008 – 2013

    Analog and Physical Design Consultant (Independent Contractor) · Cadeka

    • Executed op-amp process transfers using Cadence Virtuoso and Silvaco Gateway
    • Performed comprehensive circuit simulation and verification using Spectre and SmartSpice
    • Implemented physical designs of analog circuits with Virtuoso Layout Suite XL and Silvaco Expert
    • Ensured design integrity through thorough DRC/LVS and RC extraction
  5. 2006 – 2010

    Senior Applications Engineer · Silvaco

    • Provided expert-level support for Silvaco Gateway, SmartSpice, and SmartView CAD tools
    • Collaborated with cross-functional teams to resolve complex CAD flow challenges
    • Ran tool demonstrations for potential customers
    • As a tool vendor, trained customers in optimal use of Silvaco’s simulation and analysis tools
    • Created technical documentation and training materials for Silvaco design tools
    • Collaborated with R&D team to enhance tool capabilities based on customer feedback
    • Developed and maintained the first Silvaco Process Design Kits (PDKs) for various process nodes. Our work formed the basis of Silvaco’s current PDK group.
    • Enabled new technology node adoption through PDK development, validation, and installation
  6. 2003 – 2005

    Test Engineer · Northrop Grumman Corp

    • Updated test code for F-15 fighter jets using ATLAS programming
  7. 2000 – 2003

    Foundry Technical Representative · Sony Semiconductor

    • Utilized Cadence tools for physical design verification
    • Supported customers in implementing foundry-specific design rules
    • Provided technical guidance for process-specific design considerations
    • Provided comprehensive PDK support and maintenance for foundry customers, including help with PDK installations
    • Resolved PDK-related technical issues and implemented improvements
    • Worked with sales team to provide technical support on sales calls
  8. 1999 – 2000

    Test Engineer · TRW

    • Designed interface units to automate testing of both analog and digital VOR/ILS/Glideslope units from USAF AWACS jets
    • Wrote Test Basic code to run the required tests

Education

  1. 2026

    Post Graduate Program in AI & Machine Learning · University of Texas at Austin, McCombs

    • Coursework: ML, neural networks, computer vision, NLP, generative AI, RAG, deployment
    • Predictive maintenance: Predicted wind turbine generator failures from sensor data (~86.5% recall)
    • Generative AI: Built a RAG-based LLM medical assistant that answers clinical questions from medical manuals
    • Computer vision: Fine-tuned a VGG16 CNN to detect hard hat compliance in site images
  2. 1999

    BSEE · University of Texas San Antonio

Contact

Email is the fastest way to reach me.

williamklein@williamklein.dev
LinkedIn GitHub Download resume