At Noble Scripts, I built a production-grade RAG pipeline with Python, LlamaIndex, Qdrant, and MCP, connecting AI agents with internal databases. I also developed asynchronous authentication for a Luigi DAG pipeline and transitioned the project from a microservice to a monolith architecture.
At Epam Armenia, I processed data with PySpark and Databricks, storing results in Delta Lake and Parquet on Azure Blob Storage. I also developed streaming pipelines and worked with Kafka, Delta Live Tables, and Unity Catalog.
At Digitain, I developed a Python data collection and processing platform with APIs and a SQL Server database modeled around a Star Schema. I also developed a fraud detection pipeline and an Apache Airflow workflow for statistics and client notifications.
At I-Tech, I developed an FTP endpoint and a tool that converted unstructured PDF text into structured, SQL-ready data. My work also included a metadata extraction pipeline and an API for exposing processing operations.
