• Designed and implemented a Multi-Agent AI debugging framework using CrewAI/LangGraph to automatically diagnose Spark pipeline failures, classify errors, analyze execution plans, and generate optimized SQL/code patches. • Built autonomous remediation system capable of auto-creating Jira tickets and generating PR suggestions, reducing manual debugging effort by 60%. • Integrated LLM-driven root cause analysis with Spark metrics and SQL query plan inspection for intelligent pipeline self-healing. • Developed modular agents for log parsing, error taxonomy classification, Spark DAG analysis, and automated performance optimization.
pip install -r requirements.txt
ollama pull phi3:mini
python main.py