At Armo Tech Solutions, I designed an enterprise AI knowledge assistant that combines anomaly detection and RAG to support operational decision-making. I engineered Azure Databricks pipelines processing 150M+ records monthly and contributed to a unified anomaly scoring engine whose stakeholder adoption increased by 35%.
I also built an agentic RAG assistant with graph-based workflows, hybrid retrieval, and an automated evaluation harness. I added grounding guardrails and CI regression checks to help keep responses auditable and retrieval quality measurable.
For LLM serving, I benchmarked and optimized transformer inference with quantization, ONNX export, and continuous batching. Earlier, as a DevOps Engineer at Armo Tech Solutions, I worked on customer churn models and deployed ML services through Azure ML and AKS.

