Shubham Pardeshi

I’m a Research Analyst focused on turning large-scale data into reliable analytics and actionable AI. My work blends Python-driven data pipelines with applied machine learning for forecasting, retrieval, and automation.

At my research internship, I processed and analyzed 11+ years of ERA5-Land climate data (96,000+ hourly observations per grid location) using Python and large-scale pipelines. I also developed automated ETL workflows to generate 8,760-hour Typical Meteorological Year (TMY) datasets for climate and environmental analytics, supporting feature engineering, statistical analysis, and validation for PM2.5 forecasting and atmospheric modeling research.

In a research analyst role, I built Python-based automation pipelines for predictive analytics and large-scale data processing. I applied feature engineering, preprocessing, and model evaluation, collaborated cross-functionally, and optimized analytics workflows to automate reporting.

Across projects, I build LLM-enabled systems end to end: an AI medical research assistant integrating PubMed, OpenAlex, and ClinicalTrials APIs for retrieval across 10M+ biomedical publications, and a Social-to-Lead AI agent using LangGraph and RAG pipelines to qualify and route 1,000+ lead interactions.