Karla Brkić, Ph.D., Developer in Zagreb, Croatia
Karla is available for hire
Hire Karla

Karla Brkić, Ph.D.

Verified Expert  in Engineering

AI Researcher and Developer

Location
Zagreb, Croatia
Toptal Member Since
February 20, 2018

Karla是一名双语专业人士，拥有人工智能博士学位，在开发出色的计算机视觉方面有10多年的经验, machine learning, and AI technologies as a dedicated research scientist. As a motivational leader, 她擅长建立研究议程和管理复杂的项目，以提供世界级的产品, systems, and platforms. 她还建立了持久的关系，并利用创新思维推动前沿研究工作.

Computer VisionMachine LearningArtificial Intelligence (AI)Scientific Data AnalysisData AnalysisDeep LearningDeep Neural NetworksC++PythonJavaOpenCVPyTorchTime SeriesWekaSQLDesign PatternCryptocurrencyGameShadersVirtual RealityAugmented RealityArtificial Neural Networks

Availability

Part-time

Preferred Environment

Artificial Intelligence (AI), Computer Vision, Machine Learning, Deep Learning, Deep Neural Networks, Data Science, TensorFlow, PyTorch, Python 3, Amazon Web Services (AWS)

The most amazing...

...aspects of my career include publishing 40+ papers, consulting on projects valued up to €20M, and excelling throughout major computer vision/ML ventures.

Work Experience

Independent Expert

2019 - PRESENT
European Commission
  • Acted as a go-to resource for ML and computer vision. Assessed grant requests of up to €20M.
  • Monitored, evaluated, and advised ongoing SME projects.
  • Maintained strong alliances with various Commission units.
Technologies: IT咨询、研究、深度学习、人工智能(AI), Computer Vision, Computer Vision Algorithms

Founder | CEO

2017 - PRESENT
Inubit
  • 使用最佳实践管理提供计算机视觉/机器学习咨询服务和定制解决方案的小型企业, 在整个行业中获得主题专家的声誉.
  • Excelled throughout major projects, including researching, designing, planning, 并为AR/VR应用中使用的语义分割实施深度学习解决方案, 解决新出现的挑战，关注客户满意度.
  • 为客户建立并领导人工智能工程师团队，并利用人力资源敏锐度来提高协调和参与度.
  • 推动了一项关键计划，该计划通过深度学习方法使用体积面部重建技术.
  • 在自动加密货币交易中使用AI(深度学习、遗传编程). Delivered a separate, 提供各种见解/可视化的加密交易算法的详细性能评估.
  • 领导大规模数据分析(多个交易所数百种加密货币的价格/数量数据)，并制定了新的标准操作程序，以增加工作流程.
Technologies: OpenCV, Keras, TensorFlow，机器学习，人工智能, Artificial Neural Networks (ANN), Computer Vision, Client Success, IT Consulting, Leadership, Scikit-learn, Scikit-image, Pandas, Cryptocurrency, Blockchain, Scientific Data Analysis, Data Science, Deep Learning, Deep Neural Networks, Computer Vision Algorithms, Amazon Web Services (AWS), Google Cloud, Git, PyTorch, AutoML, Data Cleaning, Financial Modeling, SQL, Data Analysis, Data Modeling, Financial Data, Time Series, PostgreSQL, NoSQL, Docker

Independent Consultant

2016 - 2022
hamagg - bicro -克罗地亚中小企业、创新和投资机构
  • 被指定为政府机构的专家评估者，在10个以上的电话中获得价值高达700万欧元的赠款.
  • 在一般人工智能和计算机视觉主题中提供重要的见解和支持.
  • Oversaw and advised ongoing SME projects valued up to €2M.
技术领域:IT咨询、人工智能、机器学习、计算机视觉

Postdoctoral Researcher

2012 - 2017
萨格勒布大学电气工程与计算机学院
  • Applied deep learning for human detection, tracking, and privacy protection in images and videos, attaining key project milestones and producing clear, concise reporting.
  • 集成生成对抗网络，为监视系统中的隐私保护创建合成逼真的行人图像.
  • 进行其他研究，重点是检测和去识别软生物特征，同时在相关概念上支持其他研究人员. 因勤奋和卓越的职业道德而受到嘉奖.
  • Designed, trained, 并对深度神经网络在行人检测、软识别和非生物识别识别中的应用进行了评价, leveraging approaches that were scaled to other research efforts.
  • 作为一名应用认知与视觉小组的人工智能研究员，超越了自己的目标, 先进合作系统卓越研究中心.
  • 开发了一种基于Kinect点云的三维物体识别方法(c++), 促进与Vision for Robotics Group的持续沟通, Vienna University of Technology. Published two joint papers within a short timeframe.
  • Delivered other well-received initiatives; built a system for reliable pedestrian detection in videos; researched semantic video analysis and descriptors, 并在Python中生成了自动手写字符识别的快速算法.
  • 创建了一个工具，用于从Java实现的短期心率变异性段自动检测充血性心力衰竭.
Technologies: Microsoft Kinect, OpenGL, Keras, Scikit-learn, NumPy, Weka, libsvm, OpenCV, Boost, C++, Java, Python, Computer Vision, Machine Learning, Artificial Intelligence (AI), Artificial Neural Networks (ANN), Deep Learning, Deep Neural Networks, Torch, Data Cleaning, Data Analysis, Data Modeling

Visiting Researcher | Postdoc

2013 - 2014
Vienna University of Technology
  • 开发了一种从Kinect传感器获得的点云中识别3D物体的方法(c++), 促进与机器人愿景小组成员的持续沟通.
  • 获得了具有竞争力的3D物体识别性能，超越了最先进的技术.
  • Published two joint papers within a short timeframe.
技术:c++， Python, OpenCV，数据清理，数据分析，数据建模

Visiting Researcher

2011 - 2012
格拉茨工业大学电气测量与测量信号处理研究所
  • 研究了视频序列的局部外观描述符和时空语义描述. 设计了一个自定义的时空数据描述符，并用c++语言实现.
  • 采用开发的描述符，以大大减少在最先进的交通标志检测系统的错误检测的数量.
  • 与基于视觉的测量小组的成员密切合作, thriving in a multicultural environment, 在短时间内准备并在著名会议上发表多篇论文.
  • 设计并部署了详尽描述符评估的测试框架.
Technologies: Python, C++, Java, Machine Learning, Computer Vision, Artificial Intelligence (AI), Research, Scientific Data Analysis, Data Cleaning, Data Analysis, Data Modeling

Research Assistant

2008 - 2012
萨格勒布大学电气工程与计算机学院
  • 支持大规模计算机视觉研究(项目“交通控制基础设施状况测绘与评估”). 构思并实现了可靠检测和识别视频中交通标志的先进方法.
  • Researched boosting techniques, simultaneous detection of objects of different classes, and other functions, 利用每一个机会用最新的方法增加知识.
  • 使用Java构建用于自动交通标志检测和识别以及视频分析的组件, C++, and Boost; integrated the developed components with a commercial partner's existing geographic information system.
  • 创建用于数据采集和自动数据注释的系统组件.
Technologies: Subversion (SVN), libsvm, Weka, OpenCV, Boost, C++, Java, Artificial Intelligence (AI), Computer Vision, Data Cleaning, Data Analysis, Data Modeling

Semantic Segmentation for AR/VR Applications

▪AR和VR应用中语义分割的定制解决方案, based on the DeepLab v3+ architecture.
▪ Prior to development, performed a full state-of-the-art review; designed, planned, 并实现了一个实验性框架，以充分评估客户的需求并提出解决方案，同时考虑性能权衡.
▪使用TensorFlow实现自定义对象的语义分割模型，并执行数据准备, augmentation, training, and evaluation.
▪执行模型优化以实现实时性能，并将模型移植到Unity中, successfully resolving performance challenges in the process.

Volumetric Object Reconstruction From 2D Images

▪ Researched, designed, planned, 并实现了用于从2D图像重建三维物体的深度神经网络.
▪将网络集成到使用优化技术在真实场景中渲染3D物体放置的自定义解决方案中.

Data Science / ML Tools for Cryptocurrency Trading

▪为加密交易算法提供详细的性能监控系统，提供各种见解/可视化.
▪带头进行大规模数据分析(多个交易所数百种加密货币的价格/数量数据)，并制定新的标准操作程序，以增加工作流程.
▪ Applied deep learning in crypto trading.

图像和视频中个人可识别数据的去识别保护

http://www.fer.unizg.hr/demsi
▪研究通过隐藏或删除软生物特征和非生物特征标识符来保护监控视频中个人隐私的方法，同时保持数据的实用性和/或自然性.
▪ Integrated algorithms for pedestrian detection, segmentation, 以及通过使用生成对抗网络来逐帧改变外观
▪在去识别视频中启用隐写嵌入原始数据.

This was primarily built-in Python with components in C++.

Automated Detection of Congestive Heart Failure

http://www.zemris.fer.hr/~ajovic/hrzz_multisab/index.html
▪利用机器学习开发一种工具，帮助预测充血性心力衰竭.
▪该工具使用混合特征选择和旋转森林分析短期心率变异性片段.

Free Form Scanner - Automated Exam Grading

▪使用图像处理技术扫描多项选择题答题纸和自动评分的解决方案.
▪支持检测注释的回答框以及检测和识别手写的回答字母.

基于计算机视觉的交通基础设施状态映射与评估

http://www.zemris.fer.hr/~ssegvic/mastif/results_en.shtml
▪使用计算机视觉技术开发自动交通基础设施清单.
▪构思和实施先进的方法，可靠地检测和识别视频中的交通标志.
▪将开发的组件与现有的地理信息系统集成.
▪为数据采集和自动数据注释创建系统组件.

Tree Ring Analyzer for Forestry Applications

▪ Built a tool for automated tree ring detection, segmentation, 并利用计算机视觉和图像处理技术进行统计分析.
▪ Enabled performing high-precision measurements on images.

Ph. D. 论文题目:基于直方图的局部时空形态描述

http://tinyurl.com/3rdwmem5
▪ Envisioned, designed, researched, 并实现了一种在计算机视觉中表示和理解时空数据(视频数据)的新方法.
▪成功地与一个工业伙伴合作，将开发的方法用于商业交通基础设施清单中的交通标志识别, 在时间准确性方面明显优于其他可用的解决方案.
-构思、撰写并在著名国际会议上发表了十篇科学论文.
▪在多元文化环境中工作和成长，并与格拉茨理工大学保持密切合作.

Scientific Papers

http://scholar.google.hr/citations?user=i_Wl1VAAAAAJ&hl=en
我与他人合著了40多篇科学论文，发表在与计算机视觉相关的顶级同行评审会议文集和期刊上, machine learning, 和人工智能(IEEE CVPR研讨会/ IEEE会议/ Springer), Elsevier journals). See Google Scholar for a full list.

Scientific Reviewing

我曾担任顶级科学会议和期刊的审稿人，包括e.g.:
▪ IEEE Transactions on Information Forensics and Security
▪ IEEE Transactions on Biometrics, Behavior, and Identity Science
▪ IEEE Transactions on Intelligent Transportation Systems
▪ IEEE Journal of Biomedical and Health Informatics
- IEEE机器人与自动化国际会议(ICRA)
IEEE/RSJ智能机器人与系统国际会议(IROS)
- IEEE智能交通系统国际会议(ITSC)
•IEEE CVPR CV-COPS计算机视觉的光明和黑暗方面研讨会
The full list is available upon request.
2008 - 2013

Ph.D. Degree in Artificial Intelligence and Computer Vision

萨格勒布大学，电子工程与计算机学院-萨格勒布，克罗地亚

2002 - 2007

Master of Engineering in Computer Science

萨格勒布大学，电子工程与计算机学院-萨格勒布，克罗地亚

Libraries/APIs

OpenCV, TensorFlow, PyTorch, NumPy, SciPy, OpenGL, Keras, Pandas, Scikit-learn, libsvm

Tools

PyCharm, Scikit-image, LaTeX, Microsoft Visual Studio, Weka, AutoML, Git, Subversion (SVN), TensorBoard

Languages

C, C++, Python, Java, Python 3, C#, SQL, HTML

Paradigms

数据科学，测试驱动开发(TDD)，并发编程，REST，设计模式

Storage

Google Cloud, PostgreSQL, NoSQL, JSON, Redis

Frameworks

Unity3D, .NET, Unity, Boost, Microsoft Kinect, Django

Platforms

Linux，三星Gear VR, Oculus Rift，区块链，亚马逊网络服务(AWS)， Docker

Other

Facial Recognition, Computer Vision, Machine Learning, Deep Learning, Computer Science, Artificial Intelligence (AI), Image Processing, Scientific Data Analysis, Torch, Artificial Neural Networks (ANN), Client Success, IT Consulting, Research, Deep Neural Networks, Object Detection, Image Recognition, Pattern Recognition, Handwriting Recognition, Object Recognition, Data Visualization, Machine Vision, Technical Writing, Data Analysis, Data Modeling, Augmented Reality (AR), Game Design, Computer Graphics, Algorithms, HTC Vive, Virtual Reality (VR), Game Development, Cryptocurrency, Data Cleaning, Financial Modeling, Financial Data, Time Series, Shaders, Leadership, Computer Vision Algorithms, Cryptocurrency APIs, Cryptocurrency Exchanges

Collaboration That Works

How to Work with Toptal

在数小时内，而不是数周或数月，我们的网络将为您直接匹配全球行业专家.

1

Share your needs

在与Toptal领域专家的电话中讨论您的需求并细化您的范围.
2

Choose your talent

在24小时内获得专业匹配人才的简短列表，以进行审查，面试和选择.
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