Data Science for Gaming: the Anybrain’s team
In this article, we give the floor to the first number of a series of texts about data science. Data Science is essential to Anybrain’s mission of protecting all types of games on all platforms. The data we collect from human-computer interaction would be meaningless without it, and we couldn’t advise studios on the state of their games effectively.
Our data science team drives innovation in the gaming world — ensuring fair play, detecting fraud, and pioneering new ways to enhance player experiences. We recently conducted a roundtable interview with every team member, asking the same questions to learn about their personal journeys, roles, daily routines, challenges, future projects, and visions for the future of gaming.
Below, we’ve organized their insights by question, offering you a comprehensive look into the heartbeat of our team.
1. Tell Us About Yourself, Your Academic Background, and How You Joined Anybrain
Gustavo Gomes
Gustavo introduces himself with humor and candor: “I am getting old — I’ll probably be 30 already when this is published.” With a background in software engineering and studies in AI and computer graphics, he originally dropped his thesis (which didn’t feel meaningful) in favor of a gap year chasing a dream of being a professional Dota 2 player. Five years on, he’s back to finishing his thesis — this time on fraud detection using anomaly detection techniques in gaming.
Carlos Gomes
Carlos, a Computer Engineering graduate from the University of Minho, shares that his love for technology is deeply personal — enabled by his cochlear implant. A lifelong passion for gaming led him to Anybrain, where a recommendation from a colleague turned his summer internship into a full-time role. His journey is a testament to curiosity and the transformative power of technology.
Pedro Duarte
Pedro, born in Braga and holding a Master’s in Computer Engineering specializing in Intelligent Systems and Language Processing, began his career at Anybrain with a master’s thesis on esports performance. His work laid the foundation for Anybrain’s pivot toward the gaming sector, setting him on a path that now has him leading the Data Science team.
Nuno Silva
Nuno’s story begins in his hometown of Braga, where he completed an Integrated Masters in Computer Engineering at the University of Minho. A summer internship introduced him to AI/ML, and his master’s dissertation — supervised by André Pimenta (Anybrain’s CEO and co-founder) — sealed his connection with Anybrain. His passion for technology and gaming shines through in every project he undertakes.
Ana Cruz
Ana brings a unique blend of academic rigor and gaming passion. With a degree and master’s in Applied Analysis and Computation in Mathematics, she found Anybrain while exploring Portuguese startups. Excited by the opportunity to merge her analytical skills with her love for gaming, Ana joined the team to help fight cheating and ensure fair play.
2. What Is Your Specific Role Within the Team?
Gustavo Gomes
Gustavo is charged with discovering and developing new fraud detection solutions. Beyond crafting innovative models, he also monitors existing detections, performs ad hoc data quality analyses, and builds data pipelines as new challenges arise.
Carlos Gomes
Carlos focuses on creating a seamless data pipeline — from initial data preparation to training and optimizing models, and finally deploying them to catch cheaters. His role ensures that every stage of model development is executed with precision.
Pedro Duarte
As the Team Leader, Pedro coordinates the entire process: from data collection (analyzing gameplay inputs) to feature engineering, model evaluation, and even customer support. His role encompasses managing the infrastructure and guiding the team through technical challenges, ensuring that every detection meets high-quality standards.
Nuno Silva
Nuno works as a Data Scientist primarily focused on developing AI/ML models that enhance fraud detection. His contributions are key to advancing the technology behind Anybrain’s anti-cheat efforts.
Ana Cruz
Ana is part of the Data Analysis team. Her daily tasks involve ingesting and scrutinizing vast amounts of gameplay data using tools like Python, pandas, and NumPy. By applying statistical methods and machine learning algorithms, she transforms raw data into actionable insights that help identify anomalous behavior.
3. How Is a Normal Work Day for You?
Gustavo Gomes
Gustavo’s day varies considerably. Some days are spent coding — switching between R and Python for quick analyses — while other days are devoted to research and strategic thinking. He emphasizes the importance of “zooming out” to foster new ideas.
Carlos Gomes
Carlos’s routine starts with detailed data analysis before moving on to model training. He fine-tunes hyperparameters and employs interpretability tools like SHAP and LIME. Between exploring new techniques and keeping up with the latest in gaming (Silksong news included!), his day is a dynamic mix of technical and creative work.
Pedro Duarte
Pedro begins his day with a gym session, then dives into task reviews and prioritization. His schedule includes monitoring customer support, processing metrics in R/C++, leading team meetings, and strategic discussions with leadership. Evenings are reserved for research and catching up on the latest trends, balancing a demanding workload with continuous learning.
Nuno Silva
For Nuno, mornings are often dedicated to reading articles on Medium to stay updated on industry trends. He follows a structured routine that includes a gym break at lunch, intensive data preparation sessions, and model development. In his downtime, he relaxes with gaming or language classes.
Ana Cruz
Ana’s day starts by identifying a particular cheat or behavior to analyze. She then dives into research — exploring academic articles and documentation — to select a suitable method. Once she applies her chosen approach to the data, she iterates and refines her models before wrapping up the day by unwinding with some gaming, gaining firsthand insights into the player experience.
4. What Were the Biggest Problems/Barriers You Surpassed at Anybrain?
Gustavo Gomes
Gustavo highlights the inherent challenges of working with machine learning in a small company. Balancing extremely large or tiny data volumes along with computing constraints demands a great deal of tenacity — a quality he and his team have honed over time.
Carlos Gomes
Carlos reflects on both technical and personal hurdles. Initially, he had to adjust to massive datasets far beyond those on Kaggle, and he faced challenges in communication due to his cochlear implant. These experiences pushed him to grow both professionally and personally, ensuring that models accurately differentiate between normal and fraudulent behaviors.
Pedro Duarte
Pedro shares that his journey involved managing a demanding work schedule and transitioning from a one-person data role to leading an expanding team. He had to learn on the fly in an unstructured environment, particularly balancing the accuracy of fraud detection (minimizing false positives) with robust model development. Additionally, mastering the tools and infrastructure required continuous adaptation.
Nuno Silva
For Nuno, the primary challenge has been mastering organization and communication. Learning to break down and distribute his tasks effectively throughout the week has been key to navigating the complex landscape of AI/ML projects.
Ana Cruz
Ana’s biggest hurdle was transitioning from a theoretical academic background to delivering practical, actionable insights in a fast-paced environment. She also had to significantly enhance her programming skills to handle the scale and complexity of gaming data — a challenge she has met head-on.
5. Which Future Projects Do You Have in Mind?
Gustavo Gomes
Gustavo envisions a project aimed at detecting any new cheating tricks as soon as they emerge. He believes that, although it will take time to perfect, such an initiative could become a cornerstone in creating the ultimate AI-based detection system.
Carlos Gomes
Carlos is keen on exploring the deployment side of things — specifically, diving into MLOps. His future goals include learning and applying tools like Kubeflow, Kafka, Kubernetes, and Docker to ensure models are efficiently deployed and continuously updated in production environments.
Pedro Duarte
Pedro plans to expand the capabilities of internal tools already in use at Anybrain, such as the experiment tracking and fraud visualization platforms. He’s also excited about collaborating with customers with the goal of an exciting new feature (that we will share in the future), which would significantly enhance detection capabilities and tailor solutions to specific gaming environments.
Nuno Silva
Nuno doesn’t have a specific project lined up yet but is eager to deepen his knowledge of emerging technologies. He’s particularly interested in learning Spark, Kubeflow, and experimenting with Transformers to push the boundaries of what AI can do in gaming.
Ana Cruz
Ana is intrigued by the potential of quantum computing. Although still an emerging field, she sees it as a promising avenue to process and analyze data at unprecedented speeds, potentially revolutionizing cheat detection and pattern recognition in gaming.
6. How Will Data Science (or AI) Shape the Future of Gaming?
Gustavo Gomes
Gustavo believes that AI and data science will revolutionize gaming — not only by enhancing game features and competitive fairness but also by opening doors to new gameplay experiences. Imagine fair matchmaking, smart in-game assistants, adaptive economies, and even NPCs that act as integral game mechanics.
Carlos Gomes
Carlos sees data science as the key to crafting personalized, adaptive gaming experiences. By leveraging AI to adjust difficulty levels, rewards, and overall game complexity in real time, games can become more immersive and engaging — tailoring experiences to each player’s unique skill and preference.
Pedro Duarte
Pedro envisions AI as a transformative force across the industry. He predicts enhanced gameplay through adaptive challenges, more intelligent NPCs, and even streamlined content creation based on existing narratives. Furthermore, improved fraud detection ensures a secure, fair gaming environment — a balance of creativity and rigorous security.
Nuno Silva
Nuno echoes the sentiment of dynamic adaptation. He believes that as data science matures, games will continuously evolve to meet player preferences, making experiences more immersive through automatic adjustments in difficulty, rewards, and narrative elements.
Ana Cruz
Ana emphasizes that data science not only strengthens anti-cheat systems by detecting subtle behaviors but also enhances the overall gaming experience. From improved matchmaking to personalized gameplay and content development, data-driven insights will create secure environments that are engaging, and uniquely tailored to each player.
Final Thoughts
The insights from our team reveal a shared passion for gaming and technology — a blend that’s driving the future of fair play and immersive experiences. From innovative fraud detection methods to cutting-edge fields like quantum computing, the Anybrain data science team is committed to pushing boundaries and redefining what’s possible in gaming.
Other texts of this series will come out this year as we aim to showcase different aspects of our work and how our technology is shaping the present and future of video game protection.