Skilled in Python, Java, SQL, C and Unix/Linux. Experienced in distributed systems, database design, multithreading, and client-server models.
Skilled in TensorFlow, Scikit-Learn, and Pandas. Developed ensemble classifiers for sentiment analysis, regression models with outlier detection, and CNNs for object classification.
Skilled in GCP, Terraform, and Ansible. Proficient in Kubernetes, Docker, and load balancing techniques. Experienced with deployment, cloud architecture, networking, and IaC tools.
An online LaTeX editor where your universe of ideas converge. Real-time co-editing, AI-powered assistance, and unlimited collaborators — the Overleaf alternative built for teams.
Built the entire platform end-to-end as a full-stack application with Next.js and TypeScript. Designed and implemented a real-time collaborative editing system using CRDTs and WebSockets, integrated multi-provider AI assistance for writing and debugging, set up authentication and role-based access, and architected the PostgreSQL database layer with Supabase. Handled payment processing with Stripe and deployed the product through to a public waitlist launch.
Web application that uses react and springboot to emulate an application that helps people to manage projects using Scrum techniques like Jira.
Was the project manager of a team of 5 devs, which I was also part of. Did Google's course formation "Agile Project Management".
Made an High-Concept Design for a game in a team of 3 over the course of multiple sprints. Developed a playable Prototype of the concept.
The game is a combination of Vampire Survivors and Space Invaders.
Implemented finite state machines, goal-oriented behaviors (GOB), planning (GOAP) and Monte Carlo tree search (MCTS) for decision making in a game.
Developed RL agents using Tabular Q-Learning and Neural Network-based REINFORCE, training them to navigate and strategize in dynamic environments.
Developed optimized pathfinding algorithms for grid-based environments, including traditional A* and enhanced versions like Node Array A* and Gateway Heuristic.
This project explores autonomous agents navigating a dynamic environment to collect apples while managing challenges like resource scarcity and natural disasters. It evaluates different agent strategies (random, greedy, and cooperative) and their performance in various scenarios, aiming to understand agent behaviors and maximize rewards.
Used real data from E-REDES, Portugal's electricity distributor, to enhance SQL query performance by implementing indexing strategies, partitioning data, and applying SQL Server's Tuning Advisor, achieving significant efficiency gains.
Designed and deployed a multiplayer Tic-Tac-Toe web application on Google Cloud Platform, implementing full Infrastructure-as-Code. Used Terraform for provisioning compute instances and load balancers, and Ansible for configuration management and application deployment.
Developed an ensemble model to classify hotel reviews as truthful/deceptive and positive/negative with 88% accuracy. The model combines SVM for truthfulness and Naive Bayes for sentiment, using tokenization, lemmatization, and TF-IDF vectorization for preprocessing.
Analyzed real flight data from 5,623 airports, focusing on connectivity, centrality, and modularity to identify critical hubs and community structures within air travel.
Developed a 2D action-platformer from concept pitch to public-ready prototype through structured development cycles. Iterated on core combat and a dynamic three-class system based on public expo feedback, achieving positive reception from players and industry professionals.
Applied robust regression techniques, with Lasso achieving the best performance on outlier-affected data.
Built a Convolutional Neural Network (CNN) for the classification of imbalanced datasets, using data augmentation (horizontal/vertical flip, zoom, rotations) to achieve a 91% F1 score.
Benchmarked 6 IPVS schedulers in a 3-node K8s cluster (GNS3), switched kube-proxy from iptables to IPVS and measured O(1) vs O(n) lookup scaling. Project done with another colleague under guidance of the course coordinator.
Working in a six-person team, I built new Java/Spring back-end features for a Quiz Tutoring web platform and wired them into a GitLab CI pipeline that executed Spock unit, Cypress end-to-end and JMeter load tests on every commit. Writing both the code and the test suites let us ship reliable weekly releases without rollbacks.
Designed an Entity-Relationship Diagram and built a SQL Database from it. Created a web app using Python CGI scripts and HTML pages, that allows users to interact with the database.
The objective was to reduce target selection time in an abstract interface.
Achieved an 80+ SUS score for a mobile app UI by making it easy for the user.
Built a user-level File System in a Linux Environment, incorporated Data Structures, Multithreading and a Client-Server Model.
Built a Java application that serves as a manager for a distribution company. It allows to register products, customers, data from suppliers and carry out various searches on the stored information.
Built a C-based football match management system using hash tables for efficient data retrieval and precise dynamic memory allocation.
Developed a C-based logistics system using efficient arrays for product/order management, with integrated algorithms for price-based and alphabetical sorting operations.
Gonçalo Veiga
Lisbon, Portugal
gonc.veiga@gmail.com
