Autonomous AI Content Generation Engine

ABOUT

An autonomous, enterprise-grade Python content engine that discovers trending computer science and software engineering topics from 18 live internet sources, categorizes them into 19 technical domains, executes a 2-pass multi-stage LLM pipeline powered by a 5-tier resilient provider engine (Groq, Google Gemini, OpenRouter, Cohere, NVIDIA NIM), injects dynamic Mermaid.js architecture diagrams, verifies code blocks with an automated AST syntax verifier, applies an anti-AI humanizer for AdSense compliance, tracks real-time cost and latency in an observability telemetry dashboard, and publishes formatted MDX articles via GitHub Actions CI/CD.

WHAT I BUILT

• 18 Concurrent Discovery Sources parallelized in < 2.5s
• High-Throughput Algorithmic Deduplication with O(1) vector cosine similarity
• Decoupled Modular Model Registry with 5-Tier Resilient Provider Fallback
• 2-Pass Multi-Stage LLM Pipeline with Outline Caching
• Dynamic Mermaid.js Diagram Injector
• AST Code Syntax Verifier & Non-Destructive Sanitizer
• Observability & Telemetry Dashboard for token cost & latency tracking
• Automated CI/CD Workflow & Webhook Dispatch

CHALLENGES

Designing a zero-failure generation pipeline across 5 independent LLM API providers with randomized backoff jitter for HTTP 429 rate limits, while ensuring Python code blocks remain syntactically valid via AST parsing during prose sanitization.

PROJECT DETAILS

SCREENSHOTS