At Verdantas, I delivered production AI and data initiatives spanning search, agentic AI, automation, and data engineering. I built a search platform for a 300+ TB archive, helping users find files across millions of archived objects.
I also built a folder-size indexing pipeline with O(1) lookups and integrated ServiceNow APIs to turn selections into request tickets. For operational data, I developed a multi-agent system that returns ranked, evidence-based action lists across five business domains.
For Accounts Payable, I automated invoice processing with a vendor and project resolution engine, reducing the time to process 10 invoices from one hour to three minutes. I also scheduled a Fabric-to-ArcGIS pipeline that publishes cleaned and transformed data as feature layers.
My projects include a recommendation system using Apache Spark ALS and a plant-monitoring system with a PyTorch CNN that reached 98% accuracy. I work with Python, SQL, machine learning, and generative AI, including RAG and multi-agent workflows.
