Eric KnudtsonEric Knudtson

• AI Workbench is a 'one-stop shop' for ML projects designed to serve as a central organization and management hub across the MDLC. It was created to address the needs of Intuit’s data science and machine learning teams by providing a unified interface that ties together various ML tools, workflows, and project information previously scattered across the organization. AI Workbench enables managers to gain observability and insight into the hundreds of ML projects in development, production, or archived. It also fosters cross-organizational learning and collaboration by giving teams the ability to peek behind the scenes of ML projects. I worked on AI Workbench for one year, first as a solo designer with my product management partner and engineering leads from inception to launch. This included customer research and solution exploration before focusing on a specific first delivery. We expanded to include another designer and a full dev team as we gained traction.
• Prediction Service, a self-service AutoML tool designed to empower teams to train machine learning models without requiring specialized expertise. It enables users to easily experiment with a wide range of algorithms and parameters, fostering greater innovation and rapid iteration within their projects. I worked on the second version of Prediction Service for about three months. My contributions included a refactoring of the information architecture and navigation, addition of UI to enable the deployment of models, viewing and comparing experiment results, scheduling and managing automated and ad-hoc jobs, viewing model details, model history, and guardrail configuration.
• AI Launcher SDK, a Python SDK to accelerate the MDLC by reducing or eliminating “dev to prod” barriers with “portless” AI development. I was on the team for about 3 months and served as an advisor for another six months. My contributions included leading my team of one Product Manager and three Engineering leads through customer research to uncover our customers most pressing needs, followed by a Design Sprint to explore solutions and narrow on solution to prototype for the SDK - A readme on GitHub. By using the Readme Driven Development process, we iterated quickly through various versions of the SDK and made many design decisions with confidence before writing a single line of code.

• No-code authoring environment for Intuit’s Knowledge Engine, a rule-based system which drives the TurboTax experience and enables customers to receive accurate and personalized tax guidance. I joined the program for 6 months as a solo designer after it had exited the prototype and early adoption stage and was entering a phase of wider adoption. I worked with the team to design UI for authors to write tests,specify hierarchical completeness graphs, and manage deployed modules.
• No-code authoring environment for general software development via visual programming. Intuit acquired a startup that had developed a platform for creating AR experiences and IoT applications. I joined the founders, product manager, and engineering team for six months to help orient their expertise to the needs of Intuit by prototyping an authoring environment for Intuit product experiences
• Authoring tool for specifying knowledge-driven analytic optimizations

My first area of focus was on the development of technical standards and infra to enable developers across the company to work together to build a React-based UI library.

As a central Intuit Design System team we worked to shepherd components through a rigorous design process from inception to company-wide adoption. This included writing code for prototype components myself for the purpose of exploring and testing design variations before approval.

I also helped setup development infrastructure for the design system itself, including the internal website and component showcase. Today, IDS serves 11 teams with 914 designers and 5000+ developers.

On the Controls Convergence team we worked to advance the core technology, design, and user experience of control systems in industrial settings such as natural gas power plants and automobile manufacturing facilities. My individual contributions include field research at industrial plants, design of a new UI architecture, and model-based UI system to rapidly design and deploy high-performance HMI (Human-machine interfaces) in a plant automatically as machines and processes are updated or reconfigured without the need for design rework.

I also worked on the design and development of the Predix Design System for the GE Digital Industrial Internet platform, a set of reusable components and guidelines for building industrial applications. I worked with a team of designers and developers to create a system that would be used by 100s of developers across GE to build applications for the Industrial Internet.

• Nitro Snowboards (2007–2011): A European snowboarding brand offering thousands of products in seven languages.
• Nitro Photo-files (2008–2011): A Digital Asset Management (DAM) web application designed to streamline the organization and access of digital assets.
• Coal Headwear (2007–2011): A custom-built e-commerce system where I pioneered mobile-responsive web design in 2008, ensuring a seamless shopping experience across devices.