Cafe Website on AWS
•Hosted an example cafe website on an EC2 instance in multiple availability zones and connected it with an elastic load balancer and an auto scaling group to achieve high fault tolerance, high availability and scalability
•Configured security groups and network access control lists for limited access to the backend systems.
•For order processing, the EC2 instances were configured to connect to AWS Relational Database Service (RDS).
SentimentLens: Sentiment Analysis for Product Reviews
•Successfully led an MLOPS project for sentiment analysis, developing a robust NLP-based model and implementing automated deployment using Docker and Kubernetes.
•Established continuous integration and monitoring pipelines, ensuring high code quality and accurate performance tracking. The model achieved an F1 score of 0.85 and an accuracy of 87% on the test dataset, while incorporating user feedback for iterative model updates.
•Demonstrated effective collaboration with cross-functional teams, achieving seamless workflows and delivering a reliable sentiment analysis solution with improved accuracy over time.
Telecommunication Churn Prediction
•Designed a decision tree model to predict churn in a telecommunication company utilizing PySpark in Databricks.
•Preprocessed data and eliminated null values, irrelevant features and additionally executed one hot encoding on data.
•Formulated a data processing pipeline containing DecisionTree Classifier, VectorIndexer, StringIndexer which predicted overall churn rate achieving an accuracy of 86.6%.
Azure Databricks for Data Engineering Certificate from Microsoft (Coursera, 2023)