Sonia Rajput

Designed and developed a Python desktop application for SDB automation, leveraging AI-NLP technologies. Implemented custom models and NLP models such as BERT, BIOBERT, and SPACY NER to perform Named Entity Recognition (NER) and extract required entities. Trained and fine-tuned the BIOBERT model, resulting in a 50% reduction in data curation time. Collaborated with cross-functional teams to gather requirements and enhance user experience. Writing R scripts to automate data quality checks for multiple indications.

Understanding the Digit win technology. Algorithm review and improvement for real time data collection and analysis for optimized use of CNC machines. Making the algorithm robust and scalable.

Analysing Covid-19 pandemic data for Pune. Creating a Dashboard using Panel – Python for real time analysis across various wards, prabhags. R&D and Architecture design of geocoding algorithm.

Time Series Analysis and Forecasting using standard models like NNETAR, ARIMA, SARIMA, Holt-Winters, LR, MA. Built a home-grown model to provide daily, weekly, and monthly sales forecast for over 20K merchants. Reduced the error in forecast by 15% wrt state of the art models. Developed a generalised automated and intelligent inventory optimization and replenishment system for worldwide clients of #1 European Pharmaceutical Company.