Tejas Palwe

Text Classification on internal data (Sentiment Analysis):

• Developed and implemented a text classification model for sentiment analysis to assess employee satisfaction within the company.
• Built and trained a machine learning model using Keras and Tensorflow to accurately classify employee feedback into positive, negative, or neutral sentiment categories.

Text Clustering (Issues Dataset):

• Developed a text clustering in order to organizing a large corpus of Issues data into meaningful clusters.
• Used TF-IDF to transform the text data into numerical feature vectors and calculated cosine similarity to measure document similarity. Applied agglomerative clustering to group similar documents together.

Time Series Analysis:

• Developed a proof-of-concept (POC) for time series forecasting to predict the count of customer communication letters.
• Utilized Facebook's Prophet library, an open-source tool for time series analysis, to model and forecast the letter count based on historical data. Compared the results of the Prophet model with other traditional machine learning algorithms, such as ARIMA, exponential smoothing, and regression-based approaches, to assess their relative performance.
• Evaluated the performance of the Prophet model using appropriate metrics and fine-tuned the hyperparameters to optimize the forecasts.

Additional Activities:

• Conducted annual reviews of Text classification models to assess the model performance.
• Analyzed key metrics, such as accuracy, precision, recall, and F1-score, to measure the model's effectiveness.
• Interpreted predictions generated by classification models using Shap and LIME. This will help in providing insights and actionable recommendations to the business teams based on the model predictions.

• Experience in scheduling batch jobs on Control-M tool in Mainframe and Unix environment.
• Create reports using The Control-M/EM Reporting Facility.
• Basic ITIL Knowledge for process.