At GSPANN Technologies, I work on Nike’s supply-chain planning data, building and maintaining Databricks ETL pipelines with PySpark and Delta Lake. The pipelines process 10–20 GB of sales and log data.
I migrated Databricks worker nodes to AWS Graviton across 321 production tasks in six regions. Average task runtime fell by about 23%, and the migration had zero SLA breaches.
I also rebuilt the data-quality dashboards in SOLE, mapping Tableau metrics to production tables and validating parity with Tableau and direct SQL. This gave the team a single production view of data quality.
Earlier, as an AI/ML Engineer Intern at GSPANN Technologies Inc, I built an XLNet sentiment analysis model using HuggingFace and improved its accuracy from 72% to 83%. I also built a RAG document chatbot as a company internal project, using LangChain, FAISS, and a Groq LLaMA3-70b model.

