I am a passionate and motivated researcher at UNC Charlotte with background in Distributed Systems, Computer Vision and Operating Systems. I am pursuing my PhD with a focus on developing systems that solve the challenges at the Edge of the network. Currently, I work with SmartSafe StreetScapes (https://smartsafecharlotte.uncc.edu/our-team) research team, that discover ways AI can effectively improve safety and security for people and property at public spaces. My current research focuses on design and optimization of a distributed system that enables vision analytics at the Edge.
-
Experience
UNC Charlotte, Charlotte, NC (Research/Teaching Assistant: 2017-Present)
● Built an Edge test bed for latency measurement (less than 15 microsecond accuracy)
● Identified optimal keyframe selection policy for Edge video workloads that achieved 63.1% more efficiency
● Determined the best scalable Edge computing infrastructure (with 64 nodes) for real-time video analytics
● Identified 630 unique image modifications in OpenCV that result up to 90% more object detection accuracy
● Taught over 130 students over a period of 1.5 years in courses like Signal Processing, Operating Systems and Logic System Design
Robert Bosch, Bangalore, India (Software Engineer: 2014-2015)
● Development of engine management Electronic Control Unit (ECU) software for Renault and Nissan
● Configuration of interface modules CAN, LIN and EEPROM between Bosch and customer software
● Scheduling of real time OS tasks for ECU software
● Modular and overall software functionality testing for customer delivery software
● Validation for ECU software re-programmability, security and resource consumption
-
Projects
1. Distributed Messaging System for the Edge – Research Project
An adaptive messaging framework for the Edge that achieved 6.9x end-to-end latency improvement over the existing messaging systems. The messaging framework has a microservices based publish-subscribe architecture that enables communication through simple APIs.
2. Low latency Storage for the Edge – Research Project
A novel log based low latency in-memory storage architecture uniquely tailored for vision applications deployed on resource constraint Edge nodes that consume real-time video frames.
3. Adaptive Edge Latency Controller – Research Project
A latency controller that can deliver real time video frames to latency-sensitive object/event detection applications within a settling time less than 1 second. It uses Approximate Computing paradigm to simultaneously achieve application specified video frame reception latency and inference accuracy in the presence of unpredictable Wi-Fi channel interference.
4. Kernel Programming, Embedded Operating Systems – Course Projects
● JOS kernel: Developed modules that enable virtual memory, kernel and user mode, system calls, threads, context switches, interrupts, inter process communication, coordination of concurrent activities, and interface between software and hardware
● Bare-metal ARM with QEMU: Programmed bare-metal ARM with QEMU and installed embedded version of standard C library. Built Linux kernel image for ARM versatile board and a root file system with Busybox. Booted Linux with U-Boot on QEMU ARM and executed bare-metal application on QEMU ARM using U-Boot. Wrote and installed Linux kernel modules and device driver for the emulated QEMU VersatilePB board.
● Implemented configurable Cache simulator and Branch Predictors
5. Embedded Systems, IoT – Course Projects
● Implemented an IoT system that communicates sensor readings from client to HTTP and TCP servers using Tiva boards and CC3100 SimpleLink Wi-Fi module.
● Developed an embedded system that measures and communicates light intensity using Tiva boards connected through CAN bus.