Mar 2024 – May 2025 · 1 yr 2 mos · Hybrid

• 1) Developed streaming pipelines on the backend server for real-time device monitoring and workout analytics using Android, NestJS, Java, Kafka, and WebSocket, enabling prompt diagnostics and assistance.
• 2) Designed and implemented the streaming of critical data from GEMS devices to servers, addressing various edge scenarios to ensure reliability.
• 3) Led the PRISM Project, focusing on building an ML-based terrain detection system using RGB and TOF sensors. Utilized YOLO and StairNet architectures for classification and detection tasks.
• 4) Built the Group Workout feature by implementing BLE-based peer-to-peer advertisement, scanning, and GATT communication between mobile devices.
• 5) Developed BLE communication protocols between phones, leveraging the state design pattern to manage device states during GATT communication. Addressed edge cases such as temporary disconnections and BLE packet size restrictions by implementing data chunking and robust error handling.
• 6) Enabled features like creating live challenges and building dynamic networks for enhanced user interaction during workouts.

Aug 2022 – Mar 2024 · 1 yr 7 mos · Hybrid

• 1) Worked on the GEMS (Gait Enhancing And Motivating System) application, leveraging a tech stack comprising Kotlin and Android at frontend, and NestJS with TypeScript for the backend.
• 2) Took complete ownership of the design, development, and testing of the exercise data synchronization feature between the client and server. Addressed complex scenarios such as device crashes, structural differences in sensor data, missing sensor values, and GEMS device out-of-range issues. Designed and developed the complete product lifecycle, from proof of concept (POC) to a fully operational service deployed in production.
• 3) Utilized design patterns like Singleton, Observer, and Chain of Responsibility along with the MVVM architecture.
• 4) Overcame Android database constraints by implementing data compression techniques to support seamless functionality.
• 5) Worked on TizenRT, a real-time operating system (RTOS), contributing to the development of the memory_leak check tool. Designed and implemented a tool that runs memory leak detection and evaluation across the operating system, ensuring optimal memory management and performance.

Jan 2022 – Jul 2022 · 6 mos · Bengaluru, Karnataka, India · Hybrid

• 1) Hevo Activate - Reverse-ETL Platform:
• Enabled seamless movement of insights from data warehouses to CRMs like Salesforce and HubSpot.
• 2) Service Accounts for Authorization and Authentication:
• Implemented support for service accounts across the product using Google Client Library, simplifying the user onboarding process and complying with the security standards.
• 3) Retry Mechanism for Failed Transformations:
• Designed and implemented a retry mechanism for failed record transformations and egressions using Aspect-Oriented Programming and AspectJ.
• 4) Caching Module for Activate System:
• Developed a scalable caching system using Redis to store CRM database schemas, reducing redundant fetch attempts. This module was reused across multiple internal APIs and database calls.
• 5) Data Type Promotion for Amazon Redshift:
• Enabled automatic type promotion for failed records, optimizing space usage by reducing storage of failed documents.
• 6) Performance Optimization:
• Improved storage and processing efficiency by implementing file compression and parallelizing record transformations, reducing transformation time by 50%.

Jun 2021 – Jul 2021 · 1 mo · Bangalore Urban, Karnataka, India · Remote

• 1) Developed a Bixby Capsule for proactive notification tasks, enabling smart home devices to loop urgent notifications based on priority and user presence.
• 2) Designed and implemented a custom priority-based algorithm considering factors such as the user's last known location in the home, active devices, and device capabilities.
• 3) Delivered the initial working model of the project, serving as the foundation for further enhancements, including the integration of machine learning algorithms.
• 4) The solution is now deployed with SmartThings DA devices equipped with Bixby, enhancing user experience with intelligent and context-aware notifications.