Here are some projects that I've worked on independently as well as collaboratively. Enjoy!
Blueprint.ai
by UN-Solutions
UN-Solutions was the team I was involved in when developing Blueprint.at. This was our senior project where we would find a problem and use the skills and techniques we've developed to find and create a solution.
The problem we decided to tackle was productivity. When you are working on research, work and/or development you will have to sacrifice time and labor to work on presentation or report to present your findings and conclusions.
What Blueprint.ai does is give you more time to focus on your work or research by giving you a tool to quickly mock up a report with a premade layout. We also wanted this tool to be easy to use and accessible, leading us to the idea of making a web app.
To implement our idea we used Python for our API tool-set, as well as backend development. Then using Ruby for the front-end web app interface.
While I was initially tasked to work on training on a machine learning model, the team did decide to pivot to using API to handle generative text. Utilizing Huggingface tools I was able to create a summarizer module to concisely shorten text given a large document, as well another summarizer for a small description. I was tasked to test and refactor core modules that dealt with text that had to be preprocessed and fed to the ChatGPT API to generate text.
I would then take the lead in merging branches in our GitHub repository. This meant I would have to solve merge conflicts, fix security risks and add documentation where needed.
Blueprint.ai has helped me evolve as a developer. Working on a project with many cogs that are reliant upon each other. I learned how to convert criticism into a solution, discuss solutions for team conflict, reading others' code, following and creating conventions, refactoring code and finding bugs by testing the different modules.
modules worked on: <modName>, <modName>, ...
ESRB RATING PREDICTOR
BY MIGUEL A. SALVADOR T.
For this project I had to use ML techniques to train a model that is able to predict an outcome of my choosing. That would turn into using KNN model to predict ESRB ratings for games given its content.
This work required a classification model, and with time constraints I decided to use K-Nearest Neighbor. It has had a long development time and is pretty reliable. I then analyzed the dataset and preprocessed the data to optimize the model and avoid any data that can skew the modeling.
I was able to use a variety of tools used in machine learning such as pandas to analyze the dataset, numpy to store and handle values, skLearn for the base KNN model and of course kaggle to source the dataset from.
I was able to reach an accuracy of 79.6% at 24 K neighbors. Present day I would do more testing with a wider variety of models as well as do more hyperparameter tuning for the chosen model and finally add the inclusion of a more abundant dataset.
modules worked on: <modName>, <modName>, ...
DATABASE PROJECTS W/ POSTGRESS
BY MIGUEL A. SALVADOR T.
Thiis repository hosts three postgresql database projects on the modules db_01_innout_store, db_02_ipps and db_03_reviews.
In this project there is a 'convenience store' with an inventory that has to be converted into a functional database. First I am to create an Entity-Relational model with the appropriate standard utilizing crows foot notation. After the database has to be implemented with the use of Postgresql. By creating an SQL file to first write the necessary queries to then write into a Docker container containing Postgres.
In db_02 I am to normalize a large dataset from The Center for Medicare and Medicaid Services (CMS) into a PostgreSQL Database. I tackle this by using Docker as the container for my database where I create the normalized tables for the dataset. I then use Python with psycopg2 to connect with the container to insert the data. Back to the SQL file there are queries done to the database to make sure the necessary data is there.
In db_03 I am tasked to complete an interface for the user where they can checkout available rooms. The dataset is created in the SQL file where it is then used in the Docker/PostgreSQL container. I create the necessary views needed for the interface. Going back to the Py file I am testing to make sure the transactions are successful, if not I will rollback the changes and inform the user.
Both of these projects need a strong comprehension of Postgresql as well as Entity-Relational Modeling. Using Docker as a container to host Postgres database in combination with Python and Psycopg2 to connect in and out of the container to fetch and manipulate data. Security was added for both of these projects with the use of a properties file to ensure privacy, as well as using prepared statements to not only optimize queries but add a level of security against SQL injections. And lastly, a foundation of JOIN queries to find the appropriate data as efficient as possible.
modules worked on: <modName>, <modName>, ...
PROJECT NAME 4
CONTRIBUTERS
The description text will first have a portion where we will describe the project. What is the project such as description. What problem was trying to be solved. How it manages the solution process. And how it is implemented and used.
The second portion will describe my role in the project. What was my goal in the project. How I approached the issue/problem. Describe conflict/obstacles and describe my way to overcome.This is some text inside of a div block.
modules worked on: <modName>, <modName>, ...
PROJECT NAME 5
CONTRIBUTERS
The description text will first have a portion where we will describe the project. What is the project such as description. What problem was trying to be solved. How it manages the solution process. And how it is implemented and used.
The second portion will describe my role in the project. What was my goal in the project. How I approached the issue/problem. Describe conflict/obstacles and describe my way to overcome.This is some text inside of a div block.
modules worked on: <modName>, <modName>, ...
PROJECT NAME 6
CONTRIBUTERS
The description text will first have a portion where we will describe the project. What is the project such as description. What problem was trying to be solved. How it manages the solution process. And how it is implemented and used.
The second portion will describe my role in the project. What was my goal in the project. How I approached the issue/problem. Describe conflict/obstacles and describe my way to overcome.This is some text inside of a div block.
modules worked on: <modName>, <modName>, ...
