Kedar Shinde

Software Engineer

About

Hi! I'm Kedar Shinde, a third-year Computer Science undergrad. I am passionate about Algorithms, Compilers, Robotics, AI research, Computer Vision and Cellular Automatas, constantly diving into new projects and ideas. I love exploring new programming languages, especially the functional ones. I love figuring out how they offer fresh perspectives on problem-solving. I'm always eager to learn, experiment, and challenge myself with new technologies.

Experience

Research Intern - Samsung PRISM, Samsung R&D Bangalore

May 2024 - Feb 2025

Worked on End to End Speech Enhancement & Restoration using Deep Learning, where I developed a novel architecture using Autoencoders & LSTMS. Conducted rigorous experiments on noisy datasets and benchmarked models on objective metrics like PESQ and SI-SDR.

Embedded Systems Design Intern - STEMTEC AI & Robotics

April 2024 - Present

Contributed to the development of an AGV , where I was responsible for micro-ROS integration and interfacing various sensors with an STM32 board for navigation purpose.

Research Intern - Indian Summer School on Cellular Automata

May 2024 - September 2024

Under the mentorship of Dr. Ummity Srinivas Rao I researched the behaviour of 2D Cellular Automatas and the Containment Heirarchy of Subrings over Rings Z₂to Z₉ in 2D CA .

Projects

2D Cellular Automata Simulation

This project is built using Raylib and C++ allows you to run interactive simulation of 2D cellular automata, offering an intuitive GUI for creating and evolving grid-based automata systems. It allows users to draw custom patterns, adjust the grid size, and step through simulations with real-time updates. The tool has aided my research on cellular automatas and has been a real life saver.

C++ Raylib Simulation Cellular Automata

N-Body Simulation in Rust (RK4)

This Rust-based project simulates the N-body gravitational problem using the fourth-order Runge-Kutta (RK4) method for numerical integration. It efficiently models the dynamics of multiple interacting particles under Newtonian gravity and emphasizes performance, and precision.

Rust Runge-Kutta RK4 Physics Simulation Numerical Methods

SafeTV

SafeTV is a web application developed by my hackathon team, "Team Tacos" for the Smart India Hackathon selection round. It employs machine learning to detect criminal activities at railway stations using the YOLOv8 model trained on a custom dataset. The application aims to enhance security by providing real-time monitoring and alerts for suspicious behaviors in railway environments.

Python YOLOv8 Machine Learning Computer Vision Hackathon