Hi, my name is

Ricardo Silva Peres. I build things with AI.

I'm an Electrical & Computer Engineer specializing in building automated solutions powered by AI & Robotics. Currently, I'm the Head of Engineering at Geosense, I lecture at NOVA SST and conduct research on human-centric AI solutions at UNINOVA.

01.About Me

I'm an Electrical & Computer engineer with 10+ years of experience working at the intersection of Artificial Intelligence, Robotics and Data Science. I hold a PhD in Electrical & Computer Engineering and have taught 1000+ students at NOVA University of Lisbon (Portugal) since 2014. I've developed and deployed AI solutions to solve real industrial challenges for companies across Europe including Siemens A.G., Volkswagen AutoEuropa, Electrolux Professional and GKN Aerospace.

Here are a few technologies I have been working with recently:

  • Python
  • Scikit-Learn
  • PostgreSQL
  • Javascript
  • Transformers
  • MongoDB
  • Tailwind CSS
  • Pytorch
  • Docker
  • Next.js
  • OpenCV
  • ROS / ROS 2
Profile Picture

02.Where I Have Worked

  • Geosense
  • UNINOVA-CTS
  • NOVA SST
  • Introsys S.A.
  • Current Role
  • Previous Role

Head of Engineering @Geosense

Dec 2024 - Present

  • Managed the organization's software engineering projects and IT infrastructure.
  • Established and led the organization's software development team.
  • Developed the technology strategy and vision aligned with the organization's overall objectives.

03.Some Things I have Built

BSB Channel

Feature Project

Automated Youtube Shorts Channel

An automated AI video generator that uploads daily finance shorts based on scraped news. It uses a combination of Python and Selenium for scraping, a pipeline of customised BERT transformers for summarization, financial sentiment analysis and predicting social media impact and After Effects scripting for the rendering. The UI was built using Next.js, TailwindCSS and Framer Motion.

  • Python
  • Selenium
  • Transformers
  • Next.js
  • TailwindCSS
  • Framer Motion
Industrial Defect Detection

Feature Project

Structural Adhesive Inspection

Non-destructive and automated inspection of structural adhesive applications for an automotive manufacturer using RGB imagery implemented with YOLO Object Detection in Python. Since manufacturers optimize against defects, the model training was augmented with synthetic data generated with a Generative Adversarial Network (GAN). The UI was built using Dash and Plotly.

  • Python
  • GAN
  • YOLO
  • Dash
  • Plotly
  • OpenCV
BlockSense

Feature Project

Limestone Block Localization

An automated AI pipeline to localize and identify limestone blocks in quarry parks from aerial UAV imagery. It uses a combination of Python and OpenCV for image processing, several YOLO object detection models to find the IDs and finally OCR for text recognition. The UI was built using Gradio, and Folium.

  • Python
  • OpenCV
  • Gradio
  • Ultralytics
  • OCR
  • Folium

Other Noteworthy Projects

view the archive

Automatic Fruit Counting System

An AI-powered vision system that uses YOLO to count fruits from smartphone or tractor-mounted camera footage, aiding in yield estimation and harvest planning.

  • Python
  • Ultralytics
  • Object Detection

Automotive Multistage Quality Inspection

Machine learning models were used to predict downstream dimensional defects in a multistage automotive assembly line at Volkswagen's AutoEuropa plant.

  • Python
  • Pandas
  • Scikit-Learn

Non-invasive AI Diagnosis of COVID-19

A multi-modal AI system for non-invasive diagnosis of COVID-19 using a combination of sustained phonation recordings, chest X-rays and clinical history data.

  • Python
  • Scikit-Learn
  • Librosa
  • ViT

NOVAMOB: Low-cost AMR for Education

A low-cost, 3D-printed mobile robot to democratize access to hands-on robotics education world-wide. It provides features akin to commercial robots at a fraction of the cost.

  • ROS2
  • 3D Printing
  • Electronics

Sim-Based Synthetic Data for Inspection AI

A synthetic data method for automated quality inspection of structural adhesives, reducing costs and boosting model accuracy by 3.1% in data-scarce settings.

  • Python
  • CoppeliaSim
  • YOLO

Privacy-Preserving AI for Manufacturing

A federated learning framework for collaborative industrial AI, tackling data privacy and security in smart manufacturing, with a public dataset for quality inspection research.

  • Python
  • Flower
  • YOLO

04. What's Next?

Get In Touch

While I'm not actively seeking new full-time roles, my inbox is always open for new opportunities, especially consulting, fractional CTO services, or software engineering projects. If you have a question, a project in mind, or just want to say hi, feel free to reach out! I'll do my best to get back to you.

Built with Next.js, Tailwind CSS, Framer Motion, and three.js. Based on Brittany Chiang's design.
© 2024 Ricardo Silva Peres. All rights reserved.