Madhavi Pasumarthi

Built and launched CodeGlide, a platform for generating MCP servers, and developed production multi-agent systems using LangGraph, LangChain, and LiteLLM. Optimized performance to achieve an 80% reduction in token consumption and built RAG pipelines with Databricks Vector Search.

Built a distributed threat intelligence platform using Java, Python, Spark, Kafka, and Elasticsearch, and developed an ML-driven threat-detection engine that cut analyst triage time by 60%. Scaled ETL pipelines to handle 2M+ events per day and led a GCP-to-AWS migration that cut infrastructure costs by 45%.

Built backend REST and gRPC APIs and a custom Kubernetes operator for CyDrive, a secure multi-tenant cloud storage platform. Developed Kubernetes platform abstractions, controllers, sidecars, service mesh, and load balancing for production MLOps workloads.

Built Java APIs, Python services, and CLI tooling for HyScale, a developer platform that turns declarative application specs into Kubernetes deployments. Developed manifest-generation and validation workflows on Jenkins CI, cutting build times by 75%.

Pursued a Master of Science in Computer Engineering with research focus on ML robustness, classification, and prediction.

Earned a Bachelor of Engineering in Computer Engineering.

Earned the Microsoft Certified: Azure Administrator Associate certification.

Completed the Databricks Generative AI Engineering training.

Completed the Anthropic AI Fluency & Claude 101 training.

Completed the CodePath Applied AI Engineering program.

Completed the LinkedIn Learning course: Advanced Python for Agent Systems.