AI Planning Without the Slop: How I Used Agents to Build a Roadmap
FeaturedHow I used AI agents strategically—including pitting them against each other—to build a bulletproof feature roadmap without falling into the 'AI slop' trap.
I'm Cody Maughan. I bridge the gap between Enterprise Data Science and Creative Product Engineering. Currently building AI-powered video tools and medical coding solutions.
A showcase of technical products bridging AI, Blockchain, and Healthcare.

Orchestrates autonomous video production agents to compile raw text into dynamic Remotion code, bridging high-level intent with pixel-perfect execution via LangGraph and Python.

Orchestrates an AI-powered pipeline to ingest clinical notes and generate high-precision ICD-10/CPT code predictions. Leverages a FastAPI backend with a custom pgvector-based RAG architecture.

Engineers a serverless P2P social deduction arena where players deploy AI agents to deceive human opponents. Leverages PeerJS for zero-backend WebRTC synchronization and orchestrates client-side LLMs.

The 'Blue Check' for human creativity. A blockchain-based provenance protocol that allows artists to cryptographically sign their work, distinguishing human effort from AI-generated noise.
Python
TypeScript
React
Astro
LangChain
Remotion
SQL
Docker
AWS
Tailwind
Next.js
PostgreSQL
Framer Motion
OpenAI API
Gemini API
Python
TypeScript
React
Astro
LangChain
Remotion
SQL
Docker
AWS
Tailwind
Next.js
PostgreSQL
Framer Motion
OpenAI API
Gemini API
Python
TypeScript
React
Astro
LangChain
Remotion
SQL
Docker
AWS
Tailwind
Next.js
PostgreSQL
Framer Motion
OpenAI API
Gemini API How I used AI agents strategically—including pitting them against each other—to build a bulletproof feature roadmap without falling into the 'AI slop' trap.
How I took my personal site from a sluggish 56 to a perfect 100 Lighthouse score by rethinking my architecture and image strategy.
How I built an agentic video editor that writes React code instead of just generating pixels.