Jul 2024 – Oct 2024 · 3 mos · Bengaluru, Karnataka, India · Hybrid

• ◦ Test case migration: Migrated test cases from the TestNG framework to Cucumber, enabling a shift from TDD to BDD. This transformation led to improved performance, enhanced readability, and better collaboration between QA and developers.
• ◦ Dependency Management: Addressed dependency mismatches across multiple microservices, leveraging Maven to update and align dependencies. These efforts reduced system vulnerabilities and ensured seamless integration between services.
• ◦ Bug Resolution & Maintenance: Actively resolved critical bugs and tickets across microservices, ensuring the smooth functioning of the system and contributing to a stable, reliable product.

Sep 2023 – Jun 2024 · 9 mos · Bangalore Urban, Karnataka, India · On-site

• Leveraged C++ to build features that enhanced the capabilities of my project by leveraging various OOPs principles and the Wt library.
• Made architectural changes within the project where I leveraged the concept of shallow copy and deep copy allowing for a seamless flow of data across the various subcomponents within my project allowing for minimal latency and a seamless user experience.
• Migrated the backup and restoration scripts of postgres and mongo databases via ssh credentials across servers using jenkins CI/CD pipeline.
• Orchestrated the deployment of backend and frontend images using Jenkins pipelines, including build automation via Makefile, and management of Kubernetes pods updation via kubectl.
• Examined the CSV files generated with each build of the project and developed an email service that notifies the relevant user via email whenever a crash is detected.
• Worked in an agile scrum based environment.

Jun 2023 – Aug 2023 · 2 mos · Bengaluru, Karnataka, India · Hybrid

• ◦ Established Class Activation Maps: Generated class activation heatmaps highlighting areas of identification by various ImageNet models on the PETA dataset for pedestrian attribute detection.
• ◦ Performed image segmentation: Calculated IoU to find regions of overlap between class activation maps of different models on the same image.
• ◦ Predicted final label: Applied the concept of minmax game theory to predict the final class of overlapping images.
• ◦ Created a custom API : Made a custom class that allows user to create class activation maps and perform segmentation for a fed image , for multiple models.

Jun 2022 – Aug 2022 · 2 mos · Kharagpur, West Bengal, India · Remote

• Predicted the next character from a sequence of characters to ensure a user enters a secure password
• Trained a deep learning model by leveraging RNN and LSTM of 2 layers with 64 neurons each with a dropout layer of 0.2 for 15 epochs.
• The model was trained on the RockYou dataset containing over 1,000,000 leaked passwords.
• Obtained an accuracy of 70 on 1,000 passwords, 68 on 5,000 passwords and 46.74 on 100,000 passwords.