Abhishek Sharma

Created proof-of-concept to evaluate the performance of telemarketing agents using NLP. Used Google and Azure APIs to convert speech to text and analyzed the results. Performed sentiment analysis on text using the DistillBERT model and presented results by creating a dashboard from JavaScript, FastAPI, and SQL. Used Deep neural networks, LSTM, Gaussian Copula, and GANs to create synthetic time series data and tabular data. Evaluated the performance of models using multiple metrics. Worked on optimizing supply chain and clustering geospatial data using DBSCAN. Using Plotly and Streamlit, produced dashboards for real-time ECG analysis. Analyzed USA real estate data for future sales projections and created models for prediction of sales using demographics data scraped from the internet. Got introduced to graph database Neo4j, RabbitMQ, Celery, and Redis.

Collected data from web scrapping using Selenium, BeautifulSoup, and requests. Analyzed customer data and generated recommendations for promoting the companyʼs products to potential customers.

Worked on text classification and Name-Entity recognition using LSTM, spacy and Conditional random fields model and performed feature extraction for training using XML and text data. Generated recommendations using XGBoost for various customer engagements post disbursement for increasing collection/recovery and analyzed model performance using Python and Excel.