At SigmaSoft Technologies, I architected and deployed an AWS document-processing pipeline that extracts structured data from 5,000+ complex PDFs monthly and reduced manual operational costs by 70%. I also developed a Random Forest anomaly model and predictive part-assist tool that accelerated resolution time by 40%.
At IIT Patna, I developed forecasting models and fine-tuned BERT, LLaMA, and Mistral using PEFT, LoRA, and QLoRA. Earlier, at HCL Technologies, I developed Python automation for cloud data migration and ETL pipelines, and built backend services with FastAPI.

