Sree UW - Madison

About Me

I’m an undergraduate student at the University of Wisconsin – Madison, double majoring in Computer Science and Computer Engineering. I expect to graduate in Dec 2022.
I have been learning full-stack web development in the MERN (Mongo Express React Node) stack environment. I have learned numerous real-word problem-solving skills by implementing full-stack web enterprise projects. I am confident that I can make meaningful contributions to your work.
I am incredibly passionate about solving complicated, real-world problems using the power of deep learning and artificial intelligence. To broaden my knowledge, I have been doing the Deep-Learning.AI course under Dr. Andrew Ng in addition to my college course work. I expect to finish the course by December 2020. Like, Dr. Andrew said, “Artificial Intelligence is the new Electricity.”

Skills

Programing Languages: Java, C, C++, SQL, R, JavaScript, python, Dart
Frameworks: React.js, Express, Node, Next.js, GraphQl, ROS (Robot Operating System in C), MATLAB, Android Visual Studio, Flutter, Solid Work, Octave, Kubernetes (Introductory)
Databases: Mongo DB(proficient), MySQL(introductory)
API's: RestAPI
Platforms: Heroku, Vercel, Postman, Gatsby
Headless CMS: Prismic, Strapi

Research Intern

Directed Research Study under Professor Suman Bannerjee and Research Assistant Lance Hartung with team member Yingze Hou.

Main Project Description:
Augmented Reality (AR) assist goggles for firefighters. The goggles would help firefighters navigate rugged, uncertain terrain in low visibility conditions and render a 3D map of the area that other firefighters following them can use. The goggles would highlight exits and fire safety equipment like fire extinguishers in the firefighter's field of vision and save objects of interest in the firefighter's visual field.

Sub Task (Yingze Hou and I worked on):
Research in computer vision. We designed and trained a neural network from scratch for multi-class semantic segmentation and object recognition. It was inspired by the U-net model neural network. We engaged in data collection, data cleaning, model fitting, model training, model evaluation, and parameter tuning to achieve this (machine learning paradigm).

Technical Tools Utilized:

Python: The Algorithm was written in Python.
U-net: Inspired Neural Network
Pytorch: tensor library was Utilized

Engineering Intern

Project Description:
Worked alongside Dell associates in an Agile Kanban workflow setting in the Shared Component Office (SCO) of the Infrastructure Solutions Group (ISG). Built the JIRA Integration infrastructure for the SCO portal. It drastically reduced manual labor, eliminated data inconsistencies, and sped up Shared Intellectual Property (SIP) identification throughout DELL/EMC. Got real world Git/Gerrit experience and learned Functional programming and unbundled databases in a limited time frame.

Technical Tools Utilized:

Clojure (LISP): functional programming language
JIRA: backend API development
Unbundled Database: data for SCO portal was stored in unbundled database
Gerrit: for code quality control

Knights Tour Robot

Project Description:
Application Specific Integrated Circuit (ASIC). Designed and developed custom hardware for robot that would perform the knights tour on a chessboard track. The hardware was designed to minimize area and maximize response time. The designed was piplined to increase instruction throughput. The instruction to perform the Knight's tour was relayed through a bluetooth system, and the design was loaded on to a FPGA (50 MHz FPGA (Cyclone IV)).

Technical Tools Utilized:

ModelSim (Verilog): ModelSim is a multi-language environment by Mentor Graphics, for simulation of hardware description language Verilog
Mems Gyro: Utilized for maintaining robot orientation
SPI (Serial Peripheral Interface) protocol: High speed data streaming
PID (Proportional Integral Derivative) controller: For motor drive

Lane-Detection

Project Description:
Simple Algorithm for lane detection in any image and videos. Uses Python and OpenCV library to implement the functionality. Description of Algorithm: breaks down the video into frames - processes each frame individually. Converts each frame in the image into grayscale - smoothens out the image - uses Hough Transforms for getting edge image. - isolates the area of interest (lanes) - averages out the lane lines and demarcates the lane

Technical Tools Utilized:

Python: The Algorithm was written in Python.
OpenCV library: was utilized for frame by frame image manipulation

Social Network

Project Description:
Full-stack web platform that models social media platforms. Uses the MERN stack model to organize the platform. Full social media user experience with user to user chat, cookies, sessions, the user to global feed, selective access, etc.(under-development deployment phase)

Technical Tools Utilized:

React.js: Front-End UI
Redux: (NPM) Global state store
Express: Back-End business Logic
MongoDB: NoSQL database for storing user information
Node.js: Used to achieve backend implementation
Passport.js, MailChimp, Babel,etc..... : Several NPM technologies for acheiving full user experience like cookies, session, user to user chat, user to global chat, etc

Cat or not

Project Description:
Deep learning linear regression Algorithm for detecting if a given image is of a cat or not. Takes in an extensive data set of learning images and decides if the given image is of a cat or not.

Technical Tools Utilized:

Python: Algorithm was written in python
Jupyter Notebook: Algortim written in Jupyter

Activity Recognition

Project Description:
Human Activity Recognition (HAR) project under professor Umit Ogras for class ECE555. Full custom layout of a simple neural network. And post layout simulation with parasitics.

Technical Tools Utilized:

Virtuoso Layout Suite (Cadence): For custom IC layout.
Verilog: HDL for design specification

Farm Data - Analyzer

Project Description:
Java application that analyzes the milk data for several farms across the US. The application takes in milk production data of several farms as a CSV file and conducts analytics. It displays yield, tracks change across time, and generate various reports.

Technical Tools Utilized:

Java 13: Implemented Java Stream to filter and manipulate large data.
JavaFX : Used to create UI

Coding environment

Project Description:
Implemented a platformless integrated coding environment hosted on an Amazon Kubernetes cluster. Uses the enterprise implementation of Coder to access the VScode server through any web browser. It is also connected to my Github. It allows me to develop and code on any device with a web browser like iPad, phone, etc without loss of performance capabilities. Also give me an easily accessible server to ssh into.

Technical Tools Utilized:

Ubuntu: server that hosts files/ allow git connection
(Enterpirse) Code: allow coding on any browser, by using VScode server.
VSCode-server: code
AWS Kubernetes: Host server
Access control on 134.122.124.136: (level 6 password protected - invisible except for me)