* Responsible for design, implementation, testing, dial-up and launch of feature enhancements and maintaining the services that performed fraud evaluation for amazon.com and their subsidiaries world-wide* Responsible for end-end Operational Excellence tasks that involved diving deep into various microservices to find the root cause followed by a code fix* Lead the operation readiness review for peak events such as Prime day, Black Friday that involves scaling the microservice to handle over 5x of Business-As-Usual traffic, testing service resiliency and business critical levers* Launched improvements to fraud evaluation system - Inline Evaluation, significantly improving customer experience in Japan and Australia* Provided engineering support to the ML science team to prevent BIN attacks and card testing, estimated to save in the order of 1MM USD in Brazil marketplace

• Implement modern machine learning techniques to close the loop in Artificial Pancreas System (APS) Technology • Using EDA to study the behavior of Multimodal data that’s used to forecast blood glucose for Diabetes patients• Improve current algorithms for glucose forecasting accounting for other parameters such as stress, activity etc

• Used machine learning (ML) to automate the support ticket classification process for ADP's AIM Team• Created various visualization (Data exploration) to understand the underlying patterns in the data• Cleaned and preprocessed the data using CountVectorizer and Tf-Idf to transform text to numerical form• Developed an inhouse classification tool using Stochastic Gradient Descent which can perform label predictions• Deployed the tool as a web service using FastAPI to perform ML operations and return predicted categories

• Designed and developed automation tool using NX Open APIs, Visual Basic and Java to program complex and repetitive design procedures and saved more than 20% of a resource time• The automated tool ensured standardization across board reducing significant QA time and uplifted overall process quality• Scaled above methods to 8 different process improvement initiatives to increase operational efficiency and ensure quality output• Cross-trained 2 resources on NX Open APIs, Visual Basic and Java on automation methods for future needs

• Devised an automated buffer storage using Autodesk and CATIA to eliminate human intervention in assembly process• Programmed a microcontroller using C and IDP Atmel studio to initiate timely sequencing of pneumatic actuators• Lead a team of electro-mechanical and software development interns throughout the process• Was identified as a point of contact for executives to articulate challenges, timelines, and accomplishments