•Researching ways to reduce the latency of object detectors using various techniques like computation quantization, network pruning and perform time profiling of the networks on inference engines like TensorRT

• Implemented a parking lane detector for a R/C semi-truck and calibrated the global shutter camera to reduce the distortions• Designed and programmed an object detection system to avoid collision with other semi-trucks• Developed an orb-slam based script for localization and developed a simulation that used A* algorithm for path planning.• Built a software architecture with ROS as the base framework for intercommunication between four Raspberry Pi’s and three cameras to control the DC and servo motor

Part of the Data Science group. Worked on detecting and tracking artifacts in a video sequence.To predict the state of the chip and identify outliers from a high volume chip designing data

Worked with Dr. Haresh Dagale on the project " Smart walking stick for the Elderly" in the Department of Electronic Systems Engineering. The aim of the project is to detect a fall, by processing the data from the IMU sensors worn by the user. And once a fall is detected, an alert sms is sent to the registered number. Specifically my part includes implementing machine learning algorithms suitable for fall detection and implementing Linear SVM algorithm on the STM232F407 development board. The communication between activity sensors and the walking stick is done asynchronously using Bluetooth light technology.

Worked on a project titled “Porting FreeRTOS on RISC-V 1.9 Privileged Architecture” in RISE Lab under Professor Kamakoti Computer Science.Existing code of a open source real-time operating system(FreeRTOS) and spike simulator was studiedMade changes in the code of FreeRTOS which supported the latest version of RISC-V