freiberufler Senior Machine Learning Developer auf freelance.de

Senior Machine Learning Developer

offline
  • 55€/Stunde
  • 10290 Zapresic
  • auf Anfrage
  • hr  |  en
  • 23.02.2024

Kurzvorstellung

With over eight years of research and development on industrial machine learning problems, P. has experience in implementing state-of-the-art models in image detection, classification, and segmentation. When data was not labeled, he created a labe

Qualifikationen

  • Architekturvisualisierung
  • Bilderkennung
  • Data Science
  • Maschinelles Lernen
  • Objekterkennung
  • OpenCV

Projekt‐ & Berufserfahrung

Machine Learning Engineer
Good Times Tech, Zagreb
12/2023 – 12/2023 (1 Monat)
Tourismus und Freizeitwirtschaft
Tätigkeitszeitraum

12/2023 – 12/2023

Tätigkeitsbeschreibung

- Develop state of the art darts detection algorithm based on heuristic approach on limited hardware while keeping the accuracy very high.

Eingesetzte Qualifikationen

OpenCV

R&D Team Lead
Qualcomm, Zagreb
12/2020 – 12/2023 (3 Jahre, 1 Monat)
Automobilindustrie
Tätigkeitszeitraum

12/2020 – 12/2023

Tätigkeitsbeschreibung

- Developed light sword and sun glare neural network-based detectors.
- Created a small and efficient yet accurate deep-learning light source-detector.
- Created a new and improved light source classifier.
- Finished large-scale codebase handover successfully.
- Managed the annotation process with the supplier and defined a new generation annotation structure for the light source mission.
- Worked on next-generation advanced driver-assistance systems.

Eingesetzte Qualifikationen

C++, Maschinelles Lernen, Objekterkennung, OpenCV

Machine Learning Developer
Retinus Technologies, Zagreb
12/2020 – 12/2022 (2 Jahre, 1 Monat)
Innere und Äußere Sicherheit
Tätigkeitszeitraum

12/2020 – 12/2022

Tätigkeitsbeschreibung

A custom object detector in a very specific environment and I was in charge of the process that involved the object detector model, including data gathering, setting up an annotation pipeline, managing the annotation process, developing the model, and exporting the model to the iOS app.

The client didn't have any data whatsoever. The first step was data scraping and data selection. Considering the data scraping was from Google images, there were many duplicates in the dataset. I created a small annotation tool to remove duplicates, then I defined annotation instructions and set up the annotation process. I led a team of three annotators. The model development and export to the iOS platform were the final steps.

What was particularly interesting was the timeline. The project lasted for two and a half months and was successfully delivered and well received on the demo with the end client.

Eingesetzte Qualifikationen

Data Science, Objekterkennung

Research Engineer (Festanstellung)
Gideo Brothers, Zagreb
12/2019 – 12/2020 (1 Jahr, 1 Monat)
Logistikdienstleister
Tätigkeitszeitraum

12/2019 – 12/2020

Tätigkeitsbeschreibung

- Developed a depth estimation neural network with stereo video input in TensorFlow. This project included research and development of multiple state-of-the-art architectures, from Monodepth2, Struct2depth, Fast Deep Stereo to RobustMonoDE, and more.
- Created the annotation web tool in Dash/Flask, used for depth annotations.
- Implemented the neural network pipeline described in Fast Deep Stereo with 2D Convolutional Processing of cost signatures paper using TensorFlow and OpenCV.

Eingesetzte Qualifikationen

Bilderkennung, Bildverarbeitung, Maschinelles Lernen, OpenCV

Research Engineer (Festanstellung)
Microblink, Zagreb
12/2014 – 12/2019 (5 Jahre, 1 Monat)
Banken
Tätigkeitszeitraum

12/2014 – 12/2019

Tätigkeitsbeschreibung

- Developed an accurate and robust ID-1 card detector neural network that works in real-time on mobile phones, developed using TensorFlow and OpenCV.
- Developed an extremely small and accurate TensorFlow implementation of the neural network for card analysis, used for immediate user feedback.
- Built an annotation tool for detecting blur in Dash/Flask, participated in the annotation process, and developed a robust neural network classifier in TensorFlow.
- Explored and developed a face action recognizer using TensorFlow and a Visage Technologies face detector.
- Researched a Croatian ID card verification through detecting hologram using Caffe for training, and Python, OpenCV, and GIMP for data augmentation.
- Created a plugin interface in the cfSuite desktop application in C++.
Developed custom plugin creator in C++ used for the desktop cfSuite application.
- Automated application testing procedures, used for finding bugs after updates.

Eingesetzte Qualifikationen

Deeplearning4j, OpenCV

Ausbildung

MS in Computer Science
MS in Computer Science
University of Zagreb Faculty of Electrical Engineering and Computing
2017
Zagreb

Über mich

Partnering with P. means gaining access to a seasoned professional dedicated to empowering projects through the transformative power of AI. P.'s goal is to collaborate closely while leveraging AI tools as the cornerstone of innovation and success for your project.

Weitere Kenntnisse

Artificial Inteligence
Machine Learning
Computer Vision
Deep Learning
OpenCV
Data Scraping
AWS
... and more

Persönliche Daten

Sprache
  • Kroatisch (Muttersprache)
  • Englisch (Fließend)
Reisebereitschaft
auf Anfrage
Arbeitserlaubnis
  • Europäische Union
Home-Office
bevorzugt
Profilaufrufe
55
Alter
32
Berufserfahrung
9 Jahre und 10 Monate (seit 06/2014)
Projektleitung
4 Jahre

Kontaktdaten

Nur registrierte PREMIUM-Mitglieder von freelance.de können Kontaktdaten einsehen.

Jetzt Mitglied werden