PeterBucci

Led a multi-developer Agile/Scrum team in turning a real-time, audio-reactive rendering engine into a full-stack web platform for creating, saving, organizing, and launching interactive visual scenes; established the GitHub repositories, project board, sprint workflow, backlog, user stories, and code-review process.

Built core backend infrastructure with Java, Spring Boot, PostgreSQL, and Flyway, including REST APIs, authentication, scene persistence, tagging, and engagement features, with JUnit and Testcontainers integration testing against containerized PostgreSQL.

Developed and integrated React and TypeScript product features around the rendering engine, including scene creation and playback, authentication, discovery, user libraries, media controls, and reusable frontend architecture; authored 110 merged PRs across the project’s frontend, backend, and engine repositories.

Containerized and deployed the application to AWS, implementing production configuration, reverse-proxy routing, object-storage workflows, and deployment fixes to move the project from local development to a working hosted application.

Designed Recipe Lab around immutable recipe versioning: every variation becomes a new version linked to its parent, preserving the starting recipe while making ingredient, quantity, instruction, and metadata changes easy to compare. Built catalog search and filtering, recipe detail pages, lineage navigation, saves and ratings, a structured variation editor, and before-and-after recipe comparisons.

Developed the backend with Python, FastAPI, SQLAlchemy, PostgreSQL, and Alembic, including relational constraints, idempotent writes, reproducible seed data, and structured APIs. Added a secure OpenID Connect Authorization Code and PKCE account foundation with opaque application sessions, CSRF and Origin protection, onboarding, session revocation, and sensitive-data log redaction.

Built the responsive frontend with Next.js, React, and TypeScript, including a same-origin API proxy, accessible account and recipe workflows, mobile layouts, keyboard navigation, clear loading and error states, and honest shared-demo messaging during the staged transition to account-owned activity.

Created an offline recommendation and evaluation system that compares popularity, content-based, collaborative, and hybrid approaches using fixed data cutoffs, privacy guardrails, deterministic reports, and readiness gates. Also developed a curated ingredient-substitution rules engine with hard dietary and allergen constraints, explainable rankings, and explicit safety limitations.

Containerized the application with Docker Compose and built automated quality gates with GitHub Actions. Added backend, frontend, migration, accessibility, security, and full-browser acceptance coverage using pytest, Vitest, Playwright, and Axe.