Enzo

📖 Project Story

💡 Inspiration

The journey began with a simple frustration: career amnesia. As developers, we pour countless hours into our code, but when it comes time to update our resumes or portfolios, we struggle to remember what we've actually accomplished. GitHub becomes a black hole of forgotten achievements—"Did I really optimize that API latency by 40%?" or "What was that complex authentication system I built last year?"

Enzo was born from the belief that every commit tells a story, and that story should be automatically preserved and presented as a living record of professional growth. We wanted to create a platform that doesn't just track code changes, but understands their significance and transforms them into compelling career narratives.

🤖 What it does

Enzo is an Autonomous Identity Agent that bridges the gap between your engineering work and your professional presentation. It lives in your workflow and:

1. Syncs GitHub Activity: Connects to your repositories to fetch commits, PRs, and contributions in real-time.
2. Extracts "Wins" (Deep Reasoning): Uses Gemini 3 to reason over code diffs, distinguishing between trivial fixes (noise) and meaningful architectural achievements (signal).
3. Generates Content: Autonomously writes professional bios, project descriptions, and STAR-method resume bullets.
4. Builds Portfolios: Maintains a live, SEO-optimized portfolio website (enzo.stabilty.com/p/username) that updates itself.
5. Powers Career Chat (RAG): Provides an Agentic Career Coach that knows your entire history. You can ask: "Enzo, help me prep for a system design interview using my recent work on the Auth microservice."

⚙️ How we built it

Enzo is built on a modern full-stack architecture designed for the Action Era:

Frontend & Framework:

• Next.js 16 (App Router & Turbopack) for lightning-fast performance.
• TypeScript for type safety.
• Tailwind CSS + Framer Motion for a fluid, high-end developer experience.

Backend & Data:

• PostgreSQL with optimized connection pooling.
• NextAuth.js v5 handling multi-provider OAuth (GitHub, Google, LinkedIn).
• Webhook-Driven Architecture for real-time ingestion of GitHub push events.

AI & Integrations:

• Google Gemini 3 Pro: Utilizing the 1M+ token context window to ingest entire file trees and commit histories for deep reasoning.
• GitHub Apps: For secure, granular repository access.
• Dodo Payments: A live credit-based billing system that proves commercial viability.

🧠 Challenges we ran into

Building Enzo tested our assumptions about AI integration and platform scalability:

From "Summary" to "Reasoning": Early versions used smaller models that merely summarized commit messages. This failed because commit messages are often vague ("fixed bug"). The breakthrough came with Gemini 3's Long Context Reasoning. It allowed us to analyze the actual code diffs across multiple files simultaneously. This enabled Enzo to distinguish between a one-line typo fix and a critical database optimization, grounding every claim in code evidence.

GitHub Integration Complexity: GitHub Apps require careful permission management. We spent weeks debugging OAuth flows to ensure users felt confident granting repository access while maintaining strict security boundaries.

Credit System Architecture: Designing a pay-as-you-go system that felt fair required robust engineering. We implemented real-time token tracking to prevent abuse while maintaining transparency about AI usage costs.

🚀 Accomplishments that we're proud of

• Autonomous RAG System: Our AI assistant doesn't just chat—it autonomously selects search strategies to answer complex questions about users' careers.
• Real Token Transparency: Unlike most AI platforms that hide costs, Enzo shows exact token usage, building trust.
• Production-Ready: We built a fully functional SaaS with payments, auth, and database migrations in a short timeline.
• The "Living Portfolio": Seeing raw code turn into a beautiful, verified webpage automatically feels like magic.

📚 What we learned

• AI Requires Context: Small context windows limit the ability to understand professional narratives. The leap to Gemini 3's 1M token context transformed what was possible.
• Developer Tools Need Empathy: Building for developers means understanding their workflows. We learned that reliability (handling webhooks correctly) trumps flashy features.
• Billing Transparency Builds Trust: Showing real usage data, not estimates, creates confidence in the platform.

🔮 What's next for Enzo

• The Marathon Agent: Transitioning from webhook-based triggers to a fully autonomous background process that monitors codebases 24/7, proactively nudging developers when they haven't logged a 'Win' in 2 weeks.
• Multi-Modal Diagrams: Using Gemini 3 to auto-generate system architecture diagrams based on code structure changes.
• Team Features: Shared portfolios for engineering teams to track collective velocity and impact.

Built With

• agent
• autonomous
• gemini-3
• google-deepmind
• nextjs