Working as FullStack developer to develop multiple services from scratchBuilt, designed, managed and maintained multiple services from scratch involving both LLD and HLD. Some of the services which I owned fully:1. Built the backend and frontend of the Payment Service from scratch, integrating it with Stripe2. Tracking Service (integration with Aftership and EasyPost)3. Integration with Shopify to sync orders and products 4. Order Management Service to create and manage orders from our platform5. Built various important and critical features on the frontend side6. Actively participated in High Level Design of services to optimize cost and performance7. Connectivity Service for streamlining the process of shipment. It provided functionality to get rates, buy shipment, track shipment status and update shipments status based on webhook events

• Performed Unit-Testing, Integration testing and resolved various bugs in the system• Improved the centralized Query Service by documenting and resolving over a dozen error scenarios in the service and made it error-free (This service had earlier led to business loss)• Wrote monitoring scripts to send alerts across various channels for maintaining better infrastructure health• Added new REST endpoints as per the business requirements to implement new features in order to meet the product requirements• Wrote Reusable Cache and Health check libraries from scratch which can be integrated with any Java codebase for efficient use• Improved various queries which resulted in some cases up to 2X faster responses.Replaces some of the queries with transactions to ensure data consistency across different tables

Worked as a Research Intern to develop a Real-Time Human-Computer Interaction System using Hand Gestures. Responsibilities included:• Created a custom dataset of static and dynamic gestures according to the use case• Read dozens of research papers and applied ideas of some of them along with custom improvements to get the best results• Trained and tested a variety of Neural Networks (3DCNN, LSTMs) to get improved real-time performance• Developed a lightweight, low latency, deployable system that detects gestures in real-time and tested the model's performance on both Laptop's webcam as well as videos recorded using mobile phones.• Achieved categorical accuracy of 97.2% on the Test Set