Diogo Silva

SocceReach

SocceReach is a website/mobile app capable of displaying information retreived from a FIFA Football database through a premade API. The website was developed using HTML/CSS, jQuery, Knockout and JSON/AJAX. The app was made by converting the website into an APK using Adobe's PhoneGap.
This was my first large-scale University Project and first website/mobile app.

TECH: Knockout JS, jQuery, HTML5, CSS, Bootstrap4 COURSE: Introduction to Web Technologies Website

goNutri

GoNutri is a proof-of-concept website about a fictional "start-up" company offering an easy connection between customers and local, healthy restaurants. Aside from the technical aspects of the project - the website - we also had to flesh out our business idea, goals and reliability in order to impress and satisfy our companie's stakeholders. We had to create detailed reports and work in an AGILE manner with weekly SCRUMs. A fun detail about this project is that, at the time, I had no knowledge of Databases. As such, in order to allow for users to login and make purchases within the platform I used the browser's cache in order to store all necessary data and implement all of our proposed functionalities.

TECH: jQuery, HTML5, CSS, Bootstrap4 COURSE: Systems Analysis and Modeling Website, Business Start-Up, SCRUM, AGILE

SpamBot Finder

This program uses a combination of probabilistic methods such as Jaccard Distance and Simmilarity, minHash, LSH, Shingles, Sotchastic Counters and Bloom Filters in order to find Spambots in reviews left by users in the online storefront GoodOldGames. After going through the available dataset of reviews we present a list of users the program thinks could be spambots or troll reviewers by finding user pairings with very similarly worded reviews.

TEAM: Vasco Ramos TECH: Java, MatLab/Octave COURSE: Probabilistic Methods for Computer Science and Engineering Software, Similarities

Files Cleaning Helper

This project consisted in a couple of BASH scripts that allow a user (in Linux) to find files in their system and parse several options such as getting only the biggest file and it's size, ignoring files accessed after a certain date, and so on and so forth.

TEAM: Vasco Ramos TECH: Bash, Linux COURSE: Operating Systems Script

Semaphore Restaurant

The second project we had to make for the "Operating Systems" subject. On the second half of the semestre we started using the C language to develop some lesser scripts, being taught about matters such as Message Queues, MultiThreading, Deadlocks and other Process Scheduling concepts. On this project we had to, using pre-existing code given to us by our teacher, develop a "Semaphore Restaurant" where we had to control the access to the Critical Region and the management of resources between several entities running in seperate threads with the use of semaphores.

TEAM: Vasco Ramos TECH: C COURSE: Operating Systems Software, Semaphores, Multithreading

Traveling Salesperson Problem

On our course we had a subject named "Algorithms and Data Structures" that was all about data structures, optimization and algorithmic techniques (as the name implies). On our first project we had to solve the Traveling Salesperson Problem using Brute Force and Dynamic Programming, both for us to train the 2 methods and to make us experience the difference in running time between the two techniques when applied to a problem of big dimensions. Besides those 2 we also developed a Random Permutation Algorithm. Code-wise the program was developed in C and we were working with pre-made code given to us by our teacher (we pretty much just had to develop the actual algorithm to solve the problem, and not the code to actually construct it, i.e creating the cities, their distances and such). We also had to develop some Matlab/Octave scripts to both craft the graphs and help with the data (time taken to solve the problem for each technique) analysis.

TEAM: Vasco Ramos TECH:C, Matlab/Octave COURSE: Algorithms and Data Structures Software, Dynamic Programming

Binary Tree Speed Analysis

Another project we had to develop for the Algorithms and Data Structures Subject was about the analysis of the speed to perform certain operations on binary trees. We had to implement the tree's main methods (inserting a node, searching for a non-existent node and calculating it's search time, etc) alongside some Matlab/Octave scripts to help us with graphing and data analysis.

TEAM: Vasco Ramos TECH: C, Matlab/Octave COURSE: Algorithms and Data Structures Software, Binary Trees

Nomadic

My first game, first time using Unity and first time participating in a GameJam. As a team we made Nomadic in 48 hours for the Global Game Jam 2019. None of us had any knowledge of Unity or C#, yet we still had a blast learning by trial and error (mostly error on this one). Nomadic is quite the technical mess and was pretty poorly made, but as a concept I still believe it to be rather endearing and it served as a launch pad for actually learning the Unity interface and being motivated to use the engine for other personal projects.

TEAM: Pedro Oliveira , Pedro Escaleira , Mariana Graça , Tomé Azevedo TECH: Unity, C# Game, Global Game Jam

p5 Games

I tried learning the p5 framework following the tutorials and coding challenges on "The Coding Train"'s youtube channel. Most of these are challenges from those videos but there's a couple of original ones that I created to try and see where my skill with the framework was at after about a week or so.

TECH: p5 Framework, JavaScript Games

My Digital Art

I decided to get into digital drawing during the summer of 2018 just for fun and as an extra hobby. You can check out my drawings over at my reddit!

Digital Art

P5 Projects

Much like the p5 games I've also made, these were created with the help of "The Coding Train"'s youtube videos. As challenges they were pretty fun to make and I did enjoy the outcome and the experience I got from making them.

TECH: p5 Framework, JavaScript Software

myGameStack DB

This project was the first time we were tasked with constructing a database from scratch. We were first taught about Entity Relationship diagrams, then Relational Models, Relational Algebra and only finally did we learn about SQL. We were told to develop a database and an interface (using Visual Studio's C# Forms Application). We make something in the vein of IMdb and myAnimeList but targeted twoards videogames. As such, myGameStack consists of a desktop app connected to our own self-made database containing info on several games, eSports tournments, developers, publishers, and that allows for the creation of user accounts and the publishing of reviews/scores.

TEAM: Vasco Ramos TECH: SQL, C#, VisualBasic COURSE: Databases Software, Database

P2P Drive-Thru using Token Ring

On the second semester of my second year of University we had a course called "Distributed Computation" (aka Distributed Systems). On it we approached, studied and deeply followed all concepts in the book "Distributed Systems" by Andrew S. Tanenbaum and Marteen van Steen. As a first project we were told to create a "P2P Drive-Thru": A collection of several entities each running as it's own project and connected to each other in a Ring Overlay formation. All entities interacted with each other and, to manage the P2P communication/allow a process to take control and request some work/access critical resources (i.e resources that can only be used by one process at a time) we had to construct a "Token-Ring" - Using a token that continously circulates through the ring, a process gets the token, checks if the tokens empty, if so, checks if it has something to request or a message to send, puts it in the token (if it does) and passes it on to the next entity in the ring ; if the token isn't empty, the entity checks if the request on it is for itself, if so it gets the request, processes it and passes the token on, else, it simply fordwards the token.

TEAM: Vasco Ramos TECH: Python COURSE: Distributed Systems Software, Distributed Systems, P2P Communication, Token Ring

Word Counter with Map Reduce and Fault Taulerance

This project consists on a program that gets a text (or a book) and counts how many instences of each word there were through multithreading. The core of the project was centered around managing several processes counting different chunks of the text and having them synergize with each other. We used a single Coordinator, who would break the text into several partitions and distribute the work accordingly, and several Workers who would actually count the words in the chunk they were given by the coordinator using a Map Reduce Algorithm. Communication was handled using Asyncio. A vital portion of the project came from the fact that each process should be able to crash/fail (either the Coordinator or the Workers) without that affecting the final result, which we achieved through a backup coordinator and Asyncio's capabilities to detect connection breaks. As such the program is highly robust and Fault Tolerant.

TEAM: Vasco Ramos TECH: Python, Asyncio COURSE: Distributed Systems Software, Distributed Systems, P2P Communication, Map-Reduce

ChangeKeeper

One subject we had was called "Human-Computer Interactions" and it dealt with interfaces and good design practices to make a usable program/application. As a project we were simply told to think up a concept and create an app. The idea we ended up going with was a "virtual wallet" app targeted at towards day-to-day incomes/expenses and the extra functionality of managing money borrowed and lent, allowances and subscriptions. All we had to do was craft the interface using Android Studio and make it a "proof of concept", since the class wasn't really about coding, they just wanted us to make something that respected Nielson's Heuristics and was usable. However, we wanted something that we could use on a daily basis, and so we went above that and crafted a fully fledged app that is completely functional.

TEAM: Vasco Ramos TECH: Java, Android Studio COURSE: Human-Computer Interaction Software, Mobile App, Nielson's Heuristics

dimX

We had a subject called "Compilers" during our 2nd year of Uni which was all about the development and conceptualization of Programming Languages, the concepts behind formal languages and how a compiler does it's work. Our final project involved creating a language using ANTLR4 to define it (i.e a Lexer) and Java to add logic (i.e a Parser) through the use of Listeners and Visitors to walk the Syntax Tree. Our language was all about the ability the define Physical Quantities and Dimensions in order to allow for their computation (for example, a person would be able to compute the value of 10 meters divided by 2 seconds, which would output 5 m/s ; however summing 10 meters and 2 seconds would output an error). Our language would get converted to Java code, which would then be automatically compiled with the normal Java Compiler.

TEAM: Vasco Ramos , Jõao Vasconcelos , Tiago Mendes , Alina Yanchuk TECH: Java, ANTLR4 COURSE: Human-Computer Interaction Programming Language Lexer, Parser, Compiler

Resumos de Informática

This website functioned as both a compendium of all my notes, summaries and exercie resolutions and as a way to help newcomers to my course. Usually our teachers would give us work papers, tests, exercises, etc, but very rarely (or never) would they give us their resolutions or how to solve them. So I decided it'd be benefitial to give my course's (and any other who share the same subjects) freshmen my work so they'd have an easier time and wouldn't pull out as much hair as we did. I did, however, try to be careful enough as to not give them enough that would allow them to just copy over the solutions and skip doing any work (my purpose was to help and not to incentivize lazy behaviour)

TECH: HTML5, CSS, Bootstrap4, JavaScript Website

Deti4Devs 2019

This was probably the biggest project I did for AETTUA (or at least at the time it was). Deti4Devs consists in an event spanning 2 days of workshops, talks and student-company interaction taking place over at our University Department (DETI). I was both part of the Platform Team, developing a fully fledged website using Django (which I was using for the first time) to both build a REST API, File API and Interface and the Image Team for which I was tasked with creating several elements for facebook posts, posters, etc.

TEAM: AETTUA TECH: Django, Python, HTML5, CSS, Bootstrap5, JavaScript Web Platform, REST API

Bomberman AI

For our AI class we had to develop a Smart Agent capable of controlling Bomberman which would be moving around inside a recreation of Bomberman for the Nintendo Entertainment System, made in PyGame by our teacher. We were asked to use a Pathfinding algorithm (A* to be more exact) in order for our agent to be able to move around the level, kill all enemies on the screen, catch powerups and move on to the exit (which was hidden behind one of the destructable walls scattered about and could only be entered after all enemies had been killed). Levels ran from 1 to 15, with each level getting more and more varied enemies, funnily enough, however, our biggest struggle came from the starting levels where the only enemy types consisted of enemies that hugged a wall and followed it, and enemies that directly moved away from the agent (these struggles stemed from the fact that our initial bomb range without any powerups was rather limiting). Overall, pretty fun.

TECH: Python, PyGame COURSE: Artificial Intelligence AI, Pathfinding, A*, Game

Flappy Bird ML

A small project me and my friend did to pass the time. Basically, he approached me to ask if I wanted to collaborate on making a machine learning AI able to play Flappy Bird: He'd be in charge of creating the ML Algorithm, while all I had to do was make the actual game. I accepted, and we decided it'd be best if the game was made in PyGame due to Python's extensive mathematical modules which would make it easier for him to actually implement the algorithm.

TEAM: Pedro Escaleira TECH: Python, PyGame Machine Learning AI, Game

Intrusion Tracker

Intrusion Tracker consists of a web platform for the monitoring of accesses to rooms within a building. The aim was to allow admins and security staff to manage everyone who went in and out of each room, alongside allowing users to reserve rooms for meetings and such. Encompassing a backend made in Spring, a frontend made in React and a mobile app this platform was made in about 1 month.

TEAM: Vasco Ramos , Jõao Vasconcelos , Tiago Mendes TECH: Spring Framework, React JS, React Native COURSE: Introduction to Software Engineering Website, REST API, Frontend

Secure Communication in a Client-Server Architecture to Transfer Files

The goal of this project was to implement a client-server architecture with secure communication used to send files from clients to a server. Communication security was achieved through the usage of Symmetric Keys, Diffie-Hellman Key Exchange and Key Renegotiation.

TEAM: Vasco Ramos TECH: Python COURSE: Cybersecurity Software, Encryption, Cybersecurity

Secure and Authenticated Communication in a Client-Server Architecture to Transfer Files

This project was a further iteration over the previous "Secure Communication in a Client-Server Architecture" project. The goal was to improve the security of the previous implementation by adding an Authentication and Access Control step to the communication tunnel. The implementation uses a combination of Asymmetric Kyes, X.509 Certificates for server authentication, either Challenge-Response or ID Authentication for clients.

TEAM: Vasco Ramos TECH: Python COURSE: Cybersecurity Software, Encryption, Cybersecurity

Image Processing

This project consists in a program written in C that allows users to input images and process them in a variety of ways such as converting the image to and from grayscale, applying filters like sharpen and blur, adding watermarks and resizing.

TEAM: Pedro Oliveira TECH: C COURSE: Complements of Programming Languages Software, Image Processing, Data Structures

Data Compression and Video Encoding

This project consisted in a program capable of compressing and encoding text or video data in order to reduce their file size without losing information. The program was implemented per the course's specifications both in Python and C++.

TEAM: Pedro Oliveira TECH: Python, C++ COURSE: Complements of Programming Languages Software, Data Compression, Video Encoding, Golomb, OpenCV

Aether API

Aether API is a fully fledged web platform consisting of both a back and frontend capable of retreiving and showcasing air quality information to users. The goal of the course was to teach us about the importance of testing every single functionality of our programs. As such this platform was developed following a Test-Driven Development methodology where tests are written prior to actually implementing the corresponding functionalities. Both the API and frontend were tested thoroughly.

TECH: Spring Framework, SpringBoot, Java, JUnit5, Hamcrest, Mockito, Sellenium Web Driver, Awaitility, H2 Database, HTML5, CSS, Bootstrap4, JavaScript COURSE: Software Testing and Quality Control Web Platform, REST API, Unit Testing, Integration Testing, Test-Driven Development, CI/CD

Grid Marketplace

Grid Marketplace is a web platform and mobile app for buying, selling and auctioning video games. This online storefront was the second project of the Software Testing and Quality Control and had us work following an AGILE methodology with weekly SCRUMs. The web platform was built using the Spring Framework and the mobile app using React Native. The project also implemented a robust CI/CD pipeline the steps of which were shown using a complementary Python script running in a remote Raspberry PI which would inform viewers of the current deployment stage.

TEAM: Pedro Oliveira Pedro Escaleira Rafael Simões TECH: Spring Framework, SpringBoot, Java, JUnit5, Hamcrest, Mockito, Jest, Sellenium Web Driver, React Native, React JS, Raspberry Pi, PyGame COURSE: Software Testing and Quality Control Web Platform, Mobile App, REST API, Unit Testing, Integration Testing, SCRUM, AGILE, Test-Driven Development

Minerva - Social Network Mining

Minerva was the final project of my Bachelor's degree. This project consisted in a set of bots built using machine learning with the intent of infiltrating social networks (Twitter), generate controversy and insert themselves into other users. The bots mimicked human behavior, replying and creating posts in order to gather more followers. Their main goal was to get as many followers and following as many people according to a set of keywords (e.g Politics, Football, and so on). The bots would then report back with a network of connection between users which could then be showcased through our Web Platform's frontend.

TEAM: Pedro Oliveira Pedro Escaleira Rafael Simões TECH: Python, Django, React JS, PostgreSQL, Redis, Neo4J, MongoDB, ParlAI, Keras, RabbitMQ, ZeroMQ COURSE: Computer Science and Engineering Project Web Platform, REST API, Machine Learning

Traffic Sign Detection and Identification

This project was my first attempt at building and training a Machine Learning algorithm. The project itself consisted in the usage of a slew of Machine Learning methodologies for distinguishing between different Traffic Signs (GTSB). To solve this problem we used a set of Deep Neural Networks and Image Processing steps.

TEAM: Pedro Oliveira TECH: Python, Jupyter Notebook, Anaconda3, Scikit-Image, TensorFlow, Scikit-Learn, Pandas, Seaborn, Matplotlib COURSE: Machine Learning Topics Machine Learning AI, Neural Networks, LeNet, Image Processing, Computer Vision

Image Face Detection

The second Machine Learning Topics project consisted in solving the task of Face Detection in images. To tackle this issue we tested and compared several types of classifiers namely State Vector Machines, HOG Windows and Multi-Task Cascaded Convolutional Neural Networks.

TEAM: Pedro Oliveira TECH: Python, Jupyter Notebook, Anaconda3, Scikit-Image, TensorFlow, Scikit-Learn, Pandas, Seaborn, Matplotlib COURSE: Machine Learning Topics Machine Learning AI, Image Processing, Cascade Classifiers, HOG Windows w/ SVM Classifiers, MTCNN, Paper

Data Science Analysis

For this projcet we were given two distinct dataset. We were asked to analyze them through Deta Analysis steps and, according to our findings, perform Data Preparation and apply the appropriate Machine Learning Models. We were also asked to analyze the achieved results, hypothesize causes for the outputs and identify possible data mining improvements.

TEAM: Tiago Melo Pedro Pires TECH: Python, Jupyter Notebook, Anaconda3, Scikit-Image, TensorFlow, Scikit-Learn, Pandas, Seaborn, Matplotlib COURSE: Data Science Machine Learning, Data Science Analysis, Classification Algorithms, Unsupervised Learning Techniques, Data Mining

Rose Engine

The "Rose Engine" was our final project for the Computer Graphics course. Throughout the semester we had been building a slew of C++ methods for matrix computation, vector operations and Quaternion transformations. For the final project we had to use this "library" we had been building to create a propper graphical engine centered around one main scene. Our idea was to try and replicate the "Enchanted Rose" from the classic Disney movie "Beauty and the Beast". We implemented concepts such as Blinn-Phong Real Time Lighting, Shadow Mapping, Particle System, Glass Materials, Procedural Texture Generation and so on.

TEAM: Daniel Gonçalves, Henrique Gaspar, Guilherme Menezes TECH: C++, OpenGL, GLSL, GLFW COURSE: Computer Graphics Graphical Engine

Collision Avoidance Algorithms In Games

This project consisted in the implementation of several widely popular Collision Avoidance Algorithms. Namely I had to implement collision avoidance with both dynamic and static obstacles with either Priority or Blending movement. I also had to implement Collision Avoidance using the Reciprocal Velocity Obstacles algorithm.

TECH: Unity, C# COURSE: Artificial Intelligence for Games Game, AI, Collision Avoidance, Reciprocal Velocity Obstacle, Static Obstical Avoidance, Priority Movement, Blending Movement

Pathfinding Algorithms In Games

The goal of this project was to implement and compare different common Pathfinding algorithms. The algorithms included Traditional A*, NodeArray A* and JPS+. Additionally these algorithms were then integrated with Pathfollowing algorithms with Collision Avoidance.

TECH: Unity, C# COURSE: Artificial Intelligence for Games Game, AI, Pathfinding, A*, Heuristics, NodeArray A*, JPS+, Pathfollowing

Decision Making Algorithms In Games

For this project I had to implement different Decision Making algorithms. Namely Depth-Limited Goal Oriented Action Planning, Monte-Carlo Tree Search with a Limited and Unlimited Biased Playout and Rapid Action Value Estimation optimization. I also implemented an additional optimization algorithm called Child Culling which stemmed from the idea of adding world knowledge to the agent in order to more quickly discard certain branches of the decision tree. Furthermore I also created a World State representation similar to the one used in the videogame F.E.A.R. using a fixed size array with all world state data rather than the default recursive one.

TECH: Unity, C# COURSE: Artificial Intelligence for Games Game, AI, Decision Making, Depth-Limited GOAP, Monte-Carlo Tree Search, Biased Playout, Limited Biased Playout, RAVE, Child Culling, FEAR World Model

Asteroids Machine Learning AI

The concept for this project was the usage of Unity's ML-Agents framework in order to train an agent to play the classic game "Asteroids". As such, the elaboration of this work consisted in the creation of the game itself, mimicking the original's mechanics as best as possible, adapting it in order to allow for the integration with the ML-Agents framework and training agents to play the game through Reinforcement Learning. Training was done in several ways, including through the usage of Curriculum Learning with incremental stages of difficulty.

TECH: Unity, C#, Unity's MLAgents COURSE: Artificial Intelligence for Games Game, Machine Learning AI, Reinforcement Learning, PPO, Curriculum Learning

Rocket League Cooperative Bots

For this course's project we were tasked with create a set of agents who would cooperate in order to achieve a common goal. For this, we chose to create sets of bots capable of playing the videogame Rocket League. We decided to create 2 different bots. The one I ended up being in charge of was named Primus and it implemented a decision tree in order to analyze both its environment, teammates and enemies in order to decide what to do next. Rather than communicate directly with its teammates, this bot observes what they are doing and reacts accordingly. I also helped in the creation of the second group of bots - Capitão/Marujo. These bots implemented a Hive-Mind structure where two of the bots - Marujos - simply followed whatever the main bot - Capitão - told them to do. The results of either of these sets of bots was astounding as they managed to compete at a fairly high level (higher than me at least).

TEAM: Tiago Melo TECH: Python, RLBot COURSE: Autonomous Agents and Multi-Agent Systems AI, Cooperative Agents, Reactive Agents, Game, Rocket League

Diogo Gonçalves Silva

About Me

Contact Details

Project Highlights

Mafia: The Old Country

Emotionally Expressive Motion Controller

Blast Beat

Incidence

Culto

Whitted Raytracer

Skills

Work

Game Engineer

2K Games Dublin

Undergraduate Researcher

IN+ IST - Plataforma Ensino Ativo

Undergraduate Researcher

IEETA - BICenter

Education

Master's Degree

Master's in Computer Science and Engineering - Instituto Superior Técnico GPA: 19.0 / 20.0

Bachelor Degree

Bachelor's in Computer Science and Engineering - Universidade de Aveiro GPA: 18.0 / 20.0

Highschool

Ciências e Tecnologías - Secundária Dr.Manuel Laranjeira

Awards & Achievements

Published Article

Thesis Article published at IEEE ISM 2022

Academic Excellence

IST Academic Excellence Diploma

Nordic Jam Representative

GDT representative at Nordic Game Jam 2022

Best Art Direction

GJL GameParade's 2022 Award for Best Art Direction

1st Place

Think Twice 2020

CAE

Cambridge's Certificate in Advanced English

FCE

Cambridge's First Certificate of English

Scholar Merit Recognition

Rotary Club's Merit Recognition

15th Place

V National Philosophy Olympics

Honor Roll

Dr. Manuel Laranjeira School's Honor Roll

Organizations

Member

Game Dev Técnico

Member

Laboratório de Jogos

Collaborator

Nucleo de Estudantes de Informática (UA)

Member of the Board of Directors

AETTUA

Collaborator

AETTUA

Volunteer Work

Directional Assistant+

Web Summit

Referee

National Boccia Tournament

Portfolio

Mafia: The Old Country

Emotionally Expressive Motion Controller

Blast Beat

Incidence

Culto

Whitted Raytracer

Asteroids Machine Learning AI

Rocket League Cooperative Bots

Pathfinding Algorithms

Felicitá

Moirai's Trial

BotParty VR

Mafia: The Old Country

Nomadic

p5 Games

Flappy Bird ML

Neon Rush

Master's in Computer Science and Engineering - Instituto Superior Técnico
GPA: 19.0 / 20.0

Bachelor's in Computer Science and Engineering - Universidade de Aveiro
GPA: 18.0 / 20.0