Pratyush Ranjan

Cloudtuner.ai – FinOps Assistant (AI Agentic Chat-bot): Built a production-grade AI FinOps Assistant that automated cost insights and recommendations, reducing manual analysis efforts by 60%. Engineered a RAG pipeline (Python, FastAPI, LangChain, Gemini, PGVector) improving semantic retrieval accuracy by 90%+ for cloud cost and usage queries. Designed prompt orchestration and reasoning workflows enabling multi-step FinOps reporting with 30–50% faster response times. Integrated MCP servers and tool-calling, enabling real-time cost analysis and automating 20+ FinOps operations. Created a scalable, migration-safe architecture using Alembic, supporting future data models and tools with zero downtime deployments. Maigic.ai – Contributed to building a multi-AI agent ecosystem (VerifyX, VahanX, etc.) using Python, FastAPI, LangChain, AI-SDK, and ADK, improving agent task automation efficiency. Implemented short-term and long-term memory components using PostgreSQL + pgvector, increasing contextual accuracy and reducing hallucinations. Developed key VerifyX features-fraud analysis, identity verification, PAN–mobile matching, generate automated PDF reporting-boosting verification workflow s

Contributed backend features to the large-scale Total Ticketing platform, including refund, exchange, and QR/barcode scanning functionality. Built RBAC and policy-based access control for secure and granular permission management. Tech Stack: Python, Django Restframework, MySQL, Redis, Celery, GIT, AWS EC2, AWS S3

Built LifeScience CRM platform for clinical research data with Django REST + React frontend and MySQL Database. Implemented JWT auth, RBAC, Redis caching, and Celery tasks for scalability. Tech Stack: Python, Django Restframework, JavaScript, MySQL, Redis, Celery, ReactJs, GIT, BitBucket