Data Platform Development – Implemented a robust platform to streamline management of recall events for medical devices
across various clients, using Java and SpringBoot to build scalable, high-performance backend services. Leveraged PostgreSQL
database and Hibernate ORM for efficient database interactions. Developed ETL pipelines with Pandas for complex data
transformation tasks, ensuring low-latency processing of large datasets. Deployed and managed infrastructure using AWS
CloudFormation templates and AWS CDK, enabling consistent resource provisioning.
Docker and CI/CD Integration – Built and containerized applications using Docker, employing multi-stage builds to optimize
image sizes and reduce build times. Implemented CI/CD pipelines with GitHub Actions, incorporating SonarQube for static code
analysis and quality assurance. Deployed services on Kubernetes clusters using Helm charts for consistent and scalable container
orchestration. Utilized Elastic Load Balancers (ELBs) to enable fault-tolerance and horizontal scaling.
Monitoring and Performance Optimization – Implemented health monitoring for services using AWS CloudWatch, configuring
automated alarms and metrics dashboards to identify and resolve performance issues proactively. Ensured code performance
and reliability by writing unit tests using JUnit and dependency management using Maven.
Software Intern, Product Development,
BMW
June 2023 - September 2023
Python Full-Stack Development for Data Annotation (ADAS) – Developed custom web-based tools using Python (FastAPI) and
JavaScript (React) for annotating and analyzing vehicle trajectory data. These tools enabled seamless viewing and editing of
labeled log data, with features to overlay real-time sensor data on trajectory path, enabling precise analysis of how sensor inputs
affected vehicle decisions. Enabled data annotation and ‘ground truth’ labelling through both human and machine-generated labels.
Automated Testing and Validation Dashboard (ADAS) – Created a dashboard to automate the testing of path planning algorithms against a suite of predefined test cases. The dashboard facilitated simulations, captured results, and presented key performance metrics (e.g., time taken to park, distance traveled, path smoothness). Integrated automated notifications to the corresponding stakeholder/teams to quickly address any issues introduced by code changes and failure of test cases.
Collaboration and Integration – Worked closely with cross-functional teams, including UI/UX designers and data scientists, to build reusable components and integrate backend services. Facilitated seamless data flow and interaction between components, iterated quickly in an agile environment and handled integration challenges for building reliable and scalable data pipelines for machine learning development.
Machine Learning Researcher,
Game Research and Immersive Design, Rutgers
Mar 2023 - May 2023
Facial Recognition System using Machine Learning – Developed a facial emotion recognition system by developing a machine learning model based on CNNs using PyTorch. Trained the model on FER2013 dataset. Augmented the dataset using a web-scraper to gather more images. Exposed the model using FastAPI to be consumed by gaming applications. Designed and managed ML pipeline for deployment of model hosted on AWS EC2 and implemented logging in each deployment cycle.
Software Engineer,
Hewlett-Packard
Aug 2019 - May 2022
Backend System Engineering – Delivered multiple highly scalable microservice REST API backend web services using Spring framework in Java. Designed ORM mappings for business data objects with Hibernate and JPA. Implemented multithreading using Java’s ExecutorService to handle concurrent processing of large datasets, improving report generation speed by 20%. Utilized thread pools and parallel streams to optimize I/O-bound and CPU-bound tasks, enhancing throughput in generating financial and regulatory reports.
Re-Architecture of Warranty and Contract Systems – Developed an event-driven system with Java-based backend microservices in Spring Boot and Solace message queues. Leveraged multithreading through Java’s CompletableFuture and ForkJoinPool for asynchronous processing, allowing the system to handle high concurrency and meet performance demands for warranty requests on next-gen HP products. The architecture, deployed in Docker containers, was built to efficiently scale and process a large volume of parallel tasks across distributed services.
Caching for Warranty Systems – Designed and developed a scalable caching solution utilizing Redis cluster, reducing average response latency by 30%. Optimized cache updates with background refresh using ScheduledExecutorService, ensuring high availability and responsiveness even under peak load conditions.
Infrastructure Deployments – Deployed and managed Kubernetes cluster as docker containers orchestration platform for the architecture. Created pipelines for automated deployments using Jenkins. Incorporated tools such as OpenZipkin, ELK, Swagger, Kubernetes Dashboard, and Bucket4j into the architecture for distributed tracing, log management, data visualization, API documentation, auto scaling, container management, and dynamic runtime configuration.