At GEVernova, I designed a Kubernetes Operator that provisioned scoped microservices from live SCADA topology, isolating faults across 30+ sites. I also implemented Spring Boot REST APIs for UDC creation and validation, reaching 72+ requests per second while handling 10,000+ calculations.
I scaled observability across three core services with OpenTelemetry, Prometheus, and Grafana. I also developed a shared Alarm Library using Kafka to publish alarms asynchronously to a common UI, with multi-parameter filtering across 100+ applications.
For production SCADA alarm databases, I engineered a secure Model Context Protocol server for LLM-driven querying, with read-only access and query validation. I built an Initializer Service for startup configuration, DML execution, and schema loading, eliminating 90% of manual setup.
At GE Digital, I engineered a real-time C++ data pipeline that streamed 100,000+ SCADA measurements to Apache Kafka with Avro serialization and automated retries. On my PDF-QA Engine project, I implemented a Dockerized RAG pipeline for semantic search and context-aware question answering over PDF documents.
