Open to opportunities

Thiago Rateke

@thiagortk

I’m a Machine Learning and Computer Vision Engineer with a PhD, focused on turning research and ambiguous real-world problems into real products.

Brazil

Message

What I'm looking for

I'm selectively available for remote fractional and consulting engagements of 10–20 hours per week.

At Fysight, I developed computer vision and machine-learning models for BlueMirror, an interactive AI training platform for PPE and hand-hygiene practices in healthcare. I built models to recognize protective equipment and assess whether safety procedures were followed correctly.

With Luna Systems, I contributed to a monocular 3D object-detection system for motorcycle ARAS/ADAS applications. I also developed a LiDAR-free annotation approach that used monocular depth-estimation networks to generate point clouds.

At Image Processing and Computer Graphics Lab (LAPIX) in Brazilian National Institute for Digital Convergence (INCoD), I developed road-classification and segmentation methods, and created ground-truth datasets. My work included the RTK road-scene dataset, captured with low-cost cameras in traffic conditions representative of emerging countries.

Experience

Work history, roles, and key accomplishments

Current

Machine Learning Consultant

Current

Luna Systems

Feb 2021 - Present (5 years 9 months)

Contributed to monocular 3D object detection for motorcycle ARAS/ADAS applications and developed LiDAR-free annotation approaches. Built data pipelines and established computer vision practices for intelligent mobility.

Depth Estimation 3D Detection Data Annotation Data Pipelines Computer Vision Machine Learning

Current

AI Lead / Machine Learning Engineer

Current

Fysight

Aug 2020 - Present (6 years 3 months)

Developed computer vision and machine learning models for an interactive AI training platform to improve PPE and hand-hygiene practices in healthcare. Managed the complete dataset lifecycle and built real-time recognition systems for hand-hygiene steps.

Computer Vision Machine Learning Object Detection Image Classification Dataset Management Generative AI Real time video processing

Computer Vision Researcher

Image Processing and Computer Graphics Lab (LAPIX) in Brazil

Aug 2009 - Aug 2019 (10 years)

Developed road-type and road-quality classification methods with over 90% accuracy and semantic segmentation for road surfaces and damage. Created ground-truth datasets and published the RTK road-scene dataset.

Computer Vision Semantic Segmentation Convolutional Neural Networks Stereoscopy Optical Flow Dataset Creation Research

Computer Vision Engineer

Agricultural Research and Rural Extension Company of Santa C

Jun 2015 - Dec 2015 (6 months)

Maintained and stabilized a C++/OpenCV/OpenGL system for visualizing and localizing meteorological features, resolving build issues and automating dependency installation.

C++OpenCV OpenGL System Maintenance Build Automation

Software Developer

Brazillian Health Regulatory Agency (ANVISA)

Jun 2013 - Jun 2014 (1 year)

Developed an Android application and web systems supporting inspections, establishment self-assessment, administrative workflows, and structured data collection.

Android Web Development Data Collection

Education

Degrees, certifications, and relevant coursework

Federal University of Santa Catarina (UFSC)

PhD, Computer Science

2015 - 2020

PhD in Computer Science with research on visual perception and machine-learning methods for road detection, surface classification, and road-damage detection.