Ripunjoy Goswami

Project# Route Optimization for merchandisers auditing Kimberly Clark Stores in Brazil and rest of LATAM (July 2022- Present)

Roles and responsibilities:

• I worked on optimizing the routes of the Kimberly Clark merchandisers for Brazil so that they can audit the stores within their purview efficiently. I have delivered for Brazil and am extending it to Peru and Chile.
• It was challenging because the scale of the data was huge and they had laid off a certain percentage of the Merchandisers. I had to reassign the stores among the available merchandisers using distance and Store priority. Around 50 of those stores (we made sure they were low priority) based on daily and weekly merch capacity could not be covered and we intimated the business about it. Once that was reviewed with business, we ran the algorithm (had to use Pandas UDF to parallelize the code for huge data)
• Once we had the output, we loaded it into Snowflake and then was viewed on Power BI by the business. They had certain important feedback like visit duration must be a minimum of 2 hours for Club Stores and Supermarkets and 3 hours for Cash and Carry. Also, larger stores having very high sellout should mandatorily be visited on Saturdays.Implemented that successfully and now expanding to rest of LATAM.

• Project # MOPAR ETA Prediction
• Roles and responsibilities:
• I worked on MOPAR ETA while working for the automobile giant Stellantis
• I created a machine learning framework which gives us a much better expected time of arrival (ETA) for parts ordered by dealers with the company compared to legacy predictions
• Analysed the data and tried out newer algorithms to make the prediction more and more accurate
• Eventually the Extra Tree Regressor method gave us the best result
• Created a pipeline which ran every working day to give the business daily forecasts
• Once we had enough confidence in our forecast, we planned to deploy it on AWS Cloud and let go of our current legacy prediction.

• Project# Customer Sentiment Analysis using NLP Techniques
• Roles and responsibilities:
• Analyzed the customer sentiments of customers at my company by analyzing them with the help of NLP techniques
• Used techniques like Tokenization, Stemming, Lemmatization, Bag of Words, TFIDF and Word2Vec to convert text data to machine readable vectors
• Used Multinomial Naïve Bayes classifier and LSTM to classify the various sentences on a scale of 1 to 5
• (1 being the most negative and 5 being the most positive)
• Achieved this primarily with the help of TensorFlow 2.3 and Keras
• Project# Customer Segmentation using K- means Clustering
• Roles and responsibilities:
• Used Clustering analysis to segment the customer base of my company on the basis of various factors like the type of service they use, the amount of revenue they contribute, how long they have been using the services etc
• Used K- Means Clustering to carry out this analysis
• Performed Exploratory Data analysis on the source data and performed all the preprocessing steps as required
• Created clusters based on the various parameters the business wanted to identify correctly and developed a model which had a decent Silhouette Coefficient
• Deployed the final model into production once it provided an acceptable output and kept on iteratively making the changes as asked by the business
• Project # Telecom Churn Analysis using Logistic Regression
• Roles and responsibilities:
• Used the Logistic regression model to predict how many customers of my telecom company could possibly churn and move to the rival telecom providers like Verizon based on the data of the last 18 months
• Performed Exploratory Data analysis on the source data and cleaned it as required
• After that to achieve the goal of classifying the target customer base used the logistic regression model using scikit learn library of Python
• Deployed the final model into production once it provided an acceptable output and kept on iteratively making the changes as asked by the business
• Project # Visualizing Revenue Recognition Data using Tableau
• Roles and responsibilities:
• Transformed the data received from various financial billing systems and used SQL to create query on the financial results and created the relevant revenue dashboards in Tableau as per business requirements
• Brought the data in from various data sources and used an ETL tool to transform and integrate the data as per business requirements
• Dumped the final data into a database and used Tableau for reporting on it and creating dashboards as per business requirements.

• Was a part of an inhouse integration and a full implementation project in the SAP BW/ SAP BI area
• Was also a part of a couple of upgrade projects from SAP BW 7.0 to 7.3 and from 7.3 to 7.4
• Have also worked extensively on SAP HANA, SAP BODS, SAP Business Objects Suite