AHAndrew Houser.
Get in touch
Get in touch

Work & Experience

Twenty-five years
of front-end.

A track record of building design systems, leading front-end initiatives, and enabling teams to deliver exceptional user experiences.

April 2025 - Present

Team Lead, Customer Experience (CX) Domain

Milliman IntelliScript

Previously Staff Software Engineer (October 2024 - April 2025)

Leading AI implementation strategy, front-end architecture, and design system leadership for the Customer Experience (CX) domain across multiple product lines in life insurance and risk assessment.

Key initiatives

  • Design System Leadership

    Own and evolve the Sable Design System, transitioning it from a designer-led component library into a scalable, developer-first platform with reusable components and tokenized theming. Created and published blueprint projects - React SPA and Next.js starter templates that integrate the Sable Design System as a foundational layer, giving any team a well-structured starting point without gatekeeping architectural decisions.

  • Front-end Modernization

    Establish and drive front-end governance standards for architecture, accessibility, and performance across engineering teams. Modernized development tooling from Webpack to Vite to improve build performance and developer productivity. Published blueprint projects encoding architectural decisions (TanStack Query, MSW, Zustand, ESLint rules) into reusable templates that serve as both starter projects for new work and reference architectures existing teams migrate toward.

  • Accessibility & Quality

    Designed and implemented a comprehensive testing strategy using Vitest and modern testing practices, increasing coverage to 95% and reducing regression risk. Continued improving WCAG compliance, keyboard navigation, and inclusive design patterns across products.

  • Cross-Team Enablement

    Founded and lead a front-end Community of Practice to scale shared standards, documentation, and implementation patterns across teams.

  • Team Leadership

    Shape team roadmap and priorities while helping teams align on performance, maintainability, and user experience quality.

  • AI Implementation Strategy

    Leading organization-wide AI adoption strategy, defining governance frameworks, and driving cross-functional alignment on AI initiatives. Shaping enterprise AI strategy through structured evaluation processes, change management programs, and measurable implementation roadmaps that connect AI capabilities to business outcomes.

  • AI Tooling Advocacy & Workshops

    Champion AI tooling adoption within the CX team and broader engineering org. Run regular workshops covering the AI landscape, its impact on engineering and business, and practical tooling strategies - with a focus on minimizing costs and maximizing impact. Actively build and share AI agentic projects to model hands-on learning.

Selected impact

  • Standardized front-end architecture across multiple teams and applications
  • Improved developer velocity through tooling modernization and shared patterns
  • Increased product consistency and usability through design system adoption
  • Reduced production defects through improved testing and engineering standards
  • Named to the "Future of Work" council, leading cross-department AI adoption strategy

Speaking

  • “User-Centered Design in Practice” - Milliman Technology Forum, Huntington Beach, CA
  • “AI Implementation Strategies for Enterprise Engineering Teams” - AI Con USA
ReactNext.jsViteTypeScriptStorybookVitestMSWTanStackReact AriaAWSSAFe AgileGit

June 2022 - October 2024

Senior Software Engineer

TitanHST

Led development of a feature-rich admin portal focused on user experience, accessibility, and performance optimization in the healthcare sector.

Selected impact

  • Increased user engagement by 25% through a complete UX overhaul of the admin portal, creating a more intuitive interface that reduced support requests by 20%
  • Pioneered WCAG 2.1 AA compliance initiative making the platform accessible to a wider audience and increasing adoption in government and healthcare sectors
  • Optimized performance by 40% through code splitting, lazy loading, and API refactoring to improve scalability and responsiveness
  • Led customer-facing feature development, resulting in 15% increase in platform usage among key clients
ReactReduxTypeScriptBlueprintJSAWSStyled ComponentsJestStorybook

January 2020 - June 2022

Staff UI Software Engineer

Skillshare

Modernized legacy codebase, optimized performance, and mentored team members at the online learning platform serving millions of creative professionals.

Selected impact

  • Reduced feature delivery times by 30% by successfully migrating a legacy Backbone.js application to React/Next.js, enabling faster deployments and enhanced performance
  • Improved API performance by 50% through GraphQL and Apollo Client implementation, increasing content discoverability and user engagement by 20%
  • Mentored junior developers and established best practice guidelines, fostering team growth and increasing productivity by 20%
ReactNext.js 13ApolloGraphQLBootstrapMaterial UICypressJestStorybookCI/CD

July 2014 - December 2019

Senior Software Engineer

QLess

Transformed legacy systems and drove revenue growth through UX improvements for a queue management and customer engagement SaaS platform.

Selected impact

  • Boosted company revenue by $2M annually through UX/UI improvements that improved customer adoption and retention, as well as recognizing new product and up-sell opportunities
  • Modernized legacy systems transitioning from Knockout.js to modern frameworks, reducing technical debt by 50% and extending product lifecycles
  • Led cross-functional teams of 12 to deliver customer-centric solutions, reducing customer complaints by 30%
OpenLaszloKnockout.jsREST APIJavaScript

July 2002 - June 2014

Senior Product Designer / Engineer

Thomson Reuters

Developed and maintained six key products in pharmaceutical competitive intelligence, contributing to successful company acquisition and post-acquisition integration in London.

Selected impact

  • Led development and maintenance of six business-critical products that contributed to the company's successful acquisition
  • Translated complex pharmaceutical competitive intelligence data into intuitive, searchable visual interfaces, enhancing client decision-making processes
  • Played pivotal role in integrating multiple solutions into a unified user experience, merging internally developed products with acquisitions
  • Collaborated with global teams across Portland, London, and Philadelphia to ensure smooth integration and cohesive product ecosystem
ASP.NETJavaScriptXML

Side Projects

Agentic AI exploration

I learn best by building things. These projects span the full spectrum of agentic AI — from autonomous trading systems that follow global markets around the clock, to self-healing development workflows where AI detects failures and fixes code without human intervention. Together they trace the copilot-to-autopilot progression in practice.

selfHealingAI

An agentic AI system that reduces friction in the development process by adding agentic loop observers that detect changes that currently require lengthy cross-team negotiation and instead invoke AI to change code autonomously. Demonstrates self-healing patterns and reflection loops in real development workflows.

TypeScriptAI AgentsAgentic LoopsSelf-HealingReflection
  • Agentic loop observers monitor development changes and autonomously update code.
  • AI-driven code changes are invoked automatically to resolve issues without human intervention.
  • Reflection loops evaluate the quality of fixes and iterate until the system stabilizes.
  • Demonstrates the copilot-to-autopilot progression in a concrete, working system.
day-trader-ml

The evolved version of day-trader - same paper-trading agent, but I started replacing LLM decisions with trained ML models wherever pattern recognition beats language reasoning. The interesting part is figuring out where that line actually is.

PythonLightGBMFinBERTHMMFastAPINext.jsDocker
  • Trade decisions and position sizing handled by LightGBM, trained on 50+ features from real trade outcomes rather than prompted reasoning.
  • Swapped LLM sentiment scoring for FinBERT - faster, more consistent, and doesn't hallucinate market opinions.
  • Hidden Markov Model for regime detection, running in shadow mode so I can compare before committing.
  • ML models don't go live until 28 days of paired predictions look good - with a kill-switch if anything goes sideways.
day-trader

A paper-trading agent that runs 24/7, follows markets across time zones, and makes its own buy/sell decisions using an LLM. Starts with $1,000 of fake money and tries to grow it. This is where the whole thing started.

PythonOllamaFastAPIAPSchedulerNext.jsDocker
  • 19 scheduled jobs that follow the sun - Nikkei open, FTSE handoff, US session, and everything in between.
  • The LLM handles research, sentiment, trade hypotheses, and even argues the bear case against its own positions.
  • Weekly routines for rebalancing, performance review, and compacting its own memory so it doesn't lose context.
  • Full Next.js dashboard with 16 pages, proxying a FastAPI backend, all running in Docker.

Full résumé available

For a complete overview of my experience, education, and accomplishments, download my résumé or reach out directly.

Download résuméGet in touch