Guinther Kovalski da Costa, Developer in Curitiba - State of Paraná, Brazil
Guinther is available for hire
Hire Guinther

Guinther Kovalski da Costa

Verified Expert  in Engineering

Computer Vision Engineer and Developer

Location
Curitiba - State of Paraná, Brazil
Toptal Member Since
May 25, 2023

Guinther is a highly skilled computer vision engineer with 5+ years of experience building systems with millions of users. Specializing in deep learning, object detection, tracking, recognition, and algorithms, he excels at breaking complex, open-scope problems into actionable plans by determining tasks and objectives. Guinther successfully defines processing hardware, camera positioning, data requirements, system architecture, and everything involved in transforming pixels into real value.

Deep LearningComputer VisionMachine LearningArtificial Intelligence (AI)OpenCVPyTorchStatisticsScikit-learnPandasNumPyLinuxPythonDockerC++AlgorithmsPoint CloudPCL

Portfolio

PixLog
Python, TensorFlow Lite, PCL, Linear Algebra, Machine Learning...
Pacer
Linux, C++, Docker, PyTorch, OpenCV, Machine Learning, Image Recognition...
Grupo ABL
Python, Natural Language Toolkit (NLTK), Librosa, Scikit-learn, Keras...

Experience

Linux - 5 yearsOpenCV - 5 yearsComputer Vision - 5 yearsDeep Learning - 5 yearsPython - 5 yearsPyTorch - 4 yearsDocker - 3 yearsPoint Clouds - 3 years

Availability

Part-time

Preferred Environment

Linux, Python, Docker, C++, OpenCV

The most amazing...

...thing I've developed is a solution based on computer vision that optimizes the airplane boarding process and is used in 20 airports in Brazil.

Work Experience

Computer Vision Engineer

2022 - 2023
PixLog
  • Developed an algorithm to work with iPhone LiDAR and get measurement errors below 2% through advanced mathematical modeling and machine learning.
  • Improved the overall system performance by 20%, meeting the industry standards for 100x more expensive measurement systems in the market.
  • Deployed a computer vision model in very constrained hardware.
Technologies: Python, TensorFlow Lite, PCL, Linear Algebra, Machine Learning, Image Recognition, Artificial Intelligence (AI), Computer Vision Algorithms

Computer Vision Engineer

2020 - 2022
Pacer
  • Reduced the infrastructure deployment costs by 30%.
  • Created MLOps pipelines for model retraining–using Bayesian optimization–deployment, and experiment tracking.
  • Implemented the whole inference pipeline in C++ for constrained hardware.
Technologies: Linux, C++, Docker, PyTorch, OpenCV, Machine Learning, Image Recognition, Artificial Intelligence (AI), Open Neural Network Exchange (ONNX), Computer Vision Algorithms

Data Scientist

2019 - 2020
Grupo ABL
  • Automated quality assurance tasks related to hearing and searching for specific words.
  • Created and deployed a dashboard with employee id level information and multiple KPIs.
  • Built machine learning models to extract conversation topics and classify customer necessities based on brief audio descriptions.
Technologies: Python, Natural Language Toolkit (NLTK), Librosa, Scikit-learn, Keras, Machine Learning, Artificial Intelligence (AI)

Machine Learning Researcher

2018 - 2019
Groupe Renault
  • Created a proof of concept for an automatic assembly fault detection system.
  • Worked with senior manufacturing engineers in projecting the final system.
  • Published research papers with non-sensitive data and findings.
Technologies: Python, Point Clouds, 3D Scanning, Photogrammetry, Scikit-learn, Pandas, NumPy, Matplotlib, Image Recognition, Artificial Intelligence (AI), Computer Vision Algorithms

Object Detection and Distance Measurement

https://github.com/GuintherKovalski/realsense_bidings_with_yolov5
Developed a C++ library that uses Intel RealSense camera and YOLO model inference. One of the use cases of this project was automatically measuring the distance between a forklift and people in the environment, signaling the operator to slow down and even stop in case of a collision.
2020 - 2023

Master's Degree in Deep Learning

Pontifical Catholic University of Paraná - Curitiba, Paraná, Brazil

2015 - 2020

Bachelor's Degree in Robotics

Pontifical Catholic University of Paraná - Curitiba, Paraná, Brazil

SEPTEMBER 2018 - PRESENT

DL0110EN: Deep Learning with Python and PyTorch

IBM | via edX

Libraries/APIs

OpenCV, PyTorch, Scikit-learn, Pandas, NumPy, Matplotlib, Natural Language Toolkit (NLTK), Keras, PCL

Paradigms

Data Science

Platforms

Linux, Docker

Languages

C++, Python

Other

Deep Learning, Computer Vision, Machine Learning, Image Recognition, Artificial Intelligence (AI), Computer Vision Algorithms, Statistics, Linear Algebra, Image Processing, Point Clouds, 3D Scanning, Algorithms, Signal Processing, Photogrammetry, Librosa, TensorFlow Lite, Open Neural Network Exchange (ONNX)

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring