1. Data Scientist (Programmer Analyst) at Cognizant Technology Solutions
(July 2019 – May 2021)
- Explored and implemented Deep Learning techniques in the field of Natural Language Processing and Computer Vision
- Implemented end-to-end Machine Learning/Deep Learning processes such as Data Preprocessing, Exploratory Data Analysis, Data Visualization, Feature Engineering, Model Building, Evaluation and Model Deployment
- Text Summarization
- Built different iterations of Hugging Face transformers for abstractive and extractive text summarization
- Human summarization comparison was done to confirm the model abilities
- Libraries: Hugging Face, Spacy, Pandas, Streamlit, NLTK
- Rasa Chatbot
- Presented and developed Rasa conversational chatbot for FAQ use-case and highlighted advantages of Rasa X for conversation-driven development to improve efficiency by ~1.5 times
- Contributed to enhancing Rasa chatbot into multi-lingual chatbot to converse and respond with users in 6 European languages
- Libraries: Rasa, RasaX
- Optical Character Recognition
- Developed Optical Character Recognition as a service on hand-written and printed text in a tabular format using YOLO, Python tesseract and OpenCV
- Deployed using Mlflow
- Libraries: OpenCV, YOLO, Tensorflow, Mlflow
- Multi-Class Classification
- Experimentation and iteration to improve multi-class support ticket classification using data augmentation techniques and machine learning models such as Linear SVM, Multinomial Naive Bayes, Logistic Regression, Random Forest, and XGBoost
- Libraries: Sci-kit-Learn, Pandas, Numpy, Matpotlib
- Developed revenue forecasting case study for a European telecom giant using ARIMA time-series model
- Achieved an overall accuracy of ~91% with confidence interval of 95% in predicting revenue for three month rolling period
- Libraries and tools: Pandas, Numpy, Matplotlib, Python, Jupyter Notebooks, Excel