• Data Augmentation Technique has been use on images gathered from client to increase the size of the data set
• Converted pdf files into images using pdf2text and categorized images as cheques, bank statements, Remittances and invalid files.
• Built Transfer Learning VGG16 Model and Modified the last layers of pre-trained model with existing category name.
• We developed this model as a part of extracting data from cheques, remittances, bank statements and invalid files which saves time and reduces the manual process.

Developed Machine Learning model based on XGBoost for payment days category of an invoice made by the customers to find the customers who are paying in time and those who are paying late which in turn reduced the Days Sales Outstanding. The DSO was reduced by 5 days which increases the collection processby 15%. This model was deployed on AWS

• Gathering the data by selecting appropriate attributes from ORACLE SQL server
• Exploratory data analysis done to understand the insights of the data and to find missing values and outliers in the dataset and treated them with necessary techniques.
• Developed a machine learning model using clustering technique to identify the payment behavior of a customer to evaluate the risk.
• Build visualizations in tableau to present the key results for CFO’s.

• Working on the EMAGIA product for the finance based Company.
• Managed SQL database to retrieve the information which is further used to build the dashboards.
• Maintaining the Tableau server to create users, giving permission to the users for the dashboards and scheduling the dashboards refresh.
• Developed the Key CFO dashboard which was used by the CFO’s.
• Developed the dashboards for the Collections, Customer Portal Tickets and Sales Order Modules
• Developed the dashboard for Key Performance Indicator “Days Sales Outstanding”.
• Developed interactive dashboards which is used to navigate between the dashboards and implemented User access dashboards in the server using Company Id

Gathering the client requirements and relative data for the same. Performed Exploratory Data Analysis and treated missing values and outliers and performed KDD technique for the categorical attributes for label encoding. Built Machine Learning model based on Logistic Regression for one of the largest water solution company to predict the dispute category of any invoice raised by the customers. Applied hypothesis Testing and prepared visualizations in tableau.

Objective: To Classify the Reviews as Positive, Neutral and Negative reviews which helps the students to identify the good Engineering College. Extracted College details like College name, Fee, Sector, Location, Entrance exam using Beautiful soup and build the tableau dashboard which tells the basic information of engineering colleges. Extracted Reviews and Ratings for all Engineering colleges in India. Pre-processed the data using techniques like Tokenization, stop-words removal, stemming and lemmatization etc., Created word embeddings using Word2vec model and generated word cloud for the reviews. Classified the reviews as positive, neutral and negative using LSTM model.