Richa Khagwal

Built and deployed generative question-answering system for customer support, using LLM model in NLP, via PEFT & LORA techniques on text and table data. (Python, PyTorch, Deep Learning). Built semantic similarity-based search functionality using sentence embeddings in NLP chatbot.

Built and deployed BERT based text classification model for support team. Built multivariate predictive analytics models for predicting purchase sales via- LSTM, Prophet models. Fine-tuned language models like- T5, BART models for language translation.

Trained and deployed a computer vision based Indoor Navigation model for visually impaired people via- videos data frames collected manually, to perform object detection, with an F1 score of 98%. (Python, OpenCV, Tensorflow, Keras). Constructed dimension reduction techniques like- PCA, Factor Analysis to analyze the retention and graduation rate of students by department, ethnicity, and major, and reported via Tableau dashboard.

Built Univariate Time Series Forecasting models using ARIMA, ANN models for predicting monthly and weekly sales, and cross-verified results using statistical measures- RMSE, MAPE. Developed an AI conversational bot, powered with data analytics and forecasting for internal and B2B communications. (Python, NLP, Spacy, RASA chatbot)

Boosted retail sales up to 8%, by incorporating REST web services within BATA's POS system. Designed and developed a central core Master Data Management System, a web application based on Advanced Java J2EE, Spring MVC, Hibernate and performed CRUD operations on Oracle 11g database.

Developed and designed reports and built various modules on a web application in an agile software development lifecycle. Written complex SQL queries, and stored procedures to fetch data from PostgreSQL db.