Tom Osman
← Back to PortfolioProof Playbook

Proof Playbook - AI Character Platform

Yes Coach

A multi-modal AI companion platform proving UX taste, system design, and technical depth from React 19 through secured AI infrastructure on Google Cloud.

Deliverables

  • -Product strategy for AI-native coaching experiences
  • -React 19 + Vite + Tailwind application shell with fine-grained state flows
  • -Node.js + Express API backed by Firestore, Cloud Run, and Secret Manager

Core Features

  • -AI Character Management - create characters, browse trending personas, favorite them, and react with emoji micro-interactions.
  • -Multi-Modal Conversations - mix text, voice, generated images, video snippets, and real-time TTS so characters feel alive.
  • -Conversation Management - threaded chats with resume points, session history, and progress tracking.
  • -Persona System - scoped identities and guardrails per character so behaviors stay on-brief.
  • -Media & Gallery - cloud storage integration that saves uploads, model outputs, and highlight reels automatically.

Technical Architecture

  • -Frontend: React 19 + TypeScript + Vite + Tailwind for fast builds and ergonomic theming.
  • -Backend: Node.js + Express API, Firestore data models, and Cloud Run autoscaling.
  • -AI Services: Gemini, Imagen, and Veo orchestrated through a capability router.
  • -Infrastructure: Entire workload deployed on Google Cloud Platform with IaC-ready config.

Key Technical Implementations

  • -Secret Manager based API key vaulting and rotation utilities.
  • -Bidirectional WebSocket layer powering real-time voice calls.
  • -Comprehensive data modeling for characters, personas, media, and session telemetry.
  • -Signed URL media pipeline for uploads and generated assets.
  • -Complete API surface with 20+ documented endpoints.

Testing & Workflow

  • -Frontend unit + interaction coverage through Vitest + Testing Library.
  • -Backend integration suite validating every endpoint and Firestore rule.
  • -Development and deployment guides that detail branching, preview builds, and production releases.

Design & Experience

  • -UX principles that prioritize clarity, pacing, and delight for AI assistants.
  • -Code quality standards covering component boundaries, accessibility, and design tokens.

Security & Scalability

  • -Production-ready auth, audit logging, and rate-limiting considerations.
  • -Horizontal scaling strategy for chat, media, and streaming workloads.

Technical Highlights

  • -Demonstrates multi-modal AI UX, cloud infra, and thorough documentation in one proof.
Yes Coach onboarding screen showcasing AI companion features
Yes Coach - Portfolio