Portfolio Highlights
RefLog — Designing for Speed, Focus, and Flow
- Challenge: Rugby referees needed a faster, more efficient way to log match events.
- Solution: Developed RefLog, a mobile app that reduces review time from hours to minutes, emphasizing speed and usability with offline-first functionality.
- Tech: Flutter.
RefereeIQ — A Conversational AI Coach for Rugby Referees
- Challenge: Rugby referees needed a mentor-like assistant to navigate complex laws and clarifications.
- Solution: Built RefereeIQ, a conversational AI coach that provides grounded answers and clarifications, enhancing decision-making and trust.
- Tech: Flutter, OpenAI API, Firestore.
vLLM CLI UX Enhancement
- Challenge: vLLM CLI exposed 200+ configuration options with overwhelming help output.
- Solution: Designed and championed enhanced help navigation with progressive disclosure, keyword search, and group-based exploration. Merged upstream via RFC process.
- Impact: Cross-org collaboration (Red Hat + Hugging Face), merged into major open-source LLM serving framework.
- Tech: CLI UX, Miro, GitHub RFC, Open Source.
Red Hat Packages Catalog Redesign
- Challenge: Fragmented package discovery across two separate ecosystems (Trusted Libraries + AI Components) with limited filtering and poor metadata visibility.
- Solution: Designed and implemented a unified catalog with faceted filtering (7+ metadata dimensions), PatternFly 6 semantic tokens, and design version toggle for gradual rollout.
- Impact: Transformed information architecture from 2 separate pages to 1 unified catalog; improved developer experience with advanced filtering (SLSA provenance, Python version, license, release stage, category).
- Tech: React, TypeScript, PatternFly 6, Enterprise Design Systems, Information Architecture.
InstructLab Interactive Mode & Script
- Challenge: Enhance the CLI experience for developers interacting with InstructLab.
- Solution: Created a structured, user-friendly interactive mode, improving onboarding and tool adoption.
BulkPrompting Script: Streamlining CLI Bulk Input
- Challenge: Repetitive CLI workflows caused inefficiency.
- Solution: Designed and implemented a Python script for automating bulk input, improving usability.
- Tech: Python, Shell scripting.
CLI Guidelines: Designing for the Command Line
- Purpose: Establish clear, actionable guidelines for designing user-friendly and consistent CLI interfaces.
- Impact: Supports developers and UX teams in delivering seamless command-line experiences.