Pranathi Guntaka

Architected production-grade AI agentic systems for high-stakes financial workflows, reducing manual processing time by 40% and improving speed and reliability. Designed multi-turn reasoning agents integrating with internal banking tools via gRPC, delivering 30% faster verification and improved failure recovery. Deployed high-precision RAG pipelines over sensitive datasets for 50,000+ users, cutting manual research by 35%. Built LLM-powered features to process over 5TB of unstructured data and implemented LangChain-based orchestration with performance tracking across 5+ global lines, boosting development efficiency by 25%. Achieved 100ms latency targets for AI features while maintaining regulatory compliance.

Engineered high-performance backend systems using Python and Go with Apache Kafka to support scalable ML pipelines (1M+ daily requests). Optimized large-scale SQL queries and distributed schemas to boost throughput by 25%. Automated data transformation and feature engineering pipelines, reducing manual effort by 40% and reclaiming 60+ hours of monthly bandwidth. Contributed to production-grade ML engineering standards, implemented AI safety guardrails for production LLMs, and supported 99.9% system uptime with robust monitoring. Applied RAG pipelines on AWS and Kubernetes to improve data access efficiency by 30% while ensuring auditability.