Akamu E-Learning App
Akamu e.V.
(Internet und Informationstechnologie, 10-50 Mitarbeiter)
Schlagwörter
Skills
M. Sc. Autonomous Systens Master from TU Darmstadt university, previously worked as a Data Scientist at Merck KGaA.
Strong fundamentals in Machine Learning, Statistics and Linear Algebra
Technologies that I am experienced in:
Python: Numpy, PyTorch, Tensorflow, Pandas, Matplotlib, Bokeh, OpenCV, ROS, Plotly
Web: Typescript, React, REST, GraphQL, nodejs, Threejs
Database: MySQL, MongoDB
C++: OpenCV, Eigen, Qt
DevOps: Docker, Gitlab CI, CMake, git, Kubernetes
System: Linux, Bash
Strong fundamentals in Machine Learning, Statistics and Linear Algebra
Technologies that I am experienced in:
Python: Numpy, PyTorch, Tensorflow, Pandas, Matplotlib, Bokeh, OpenCV, ROS, Plotly
Web: Typescript, React, REST, GraphQL, nodejs, Threejs
Database: MySQL, MongoDB
C++: OpenCV, Eigen, Qt
DevOps: Docker, Gitlab CI, CMake, git, Kubernetes
System: Linux, Bash
Projekthistorie
08/2017
-
bis jetzt
The Akamu e-learning app, similar to quiz duell, allows students to play against each other by scoring points by answering difficult questions on the lectures they are currently taking.
I took on the role of a fullstack developer and were responsible for the backend development in golang and the administration webapp for the content creators in angularjs.
Over the last years my role shifted towards supervising and teaching new members of our student organisation that further push the development of the Akamu app.
I took on the role of a fullstack developer and were responsible for the backend development in golang and the administration webapp for the content creators in angularjs.
Over the last years my role shifted towards supervising and teaching new members of our student organisation that further push the development of the Akamu app.
07/2021
-
10/2022
Computer Vision Machine Learning Engineer
(Konsumgüter und Handel, < 10 Mitarbeiter)
Early stage startup MVP development.
Task was to build a system that detects and recognizes grocaries.
Setup the collection and labelling of the data.
Designed and trained CNN detector that detects and classifies different features on products, (Name, EAN, Barcode)
Parse detections and pipe them into an OCR.
Implemented weighted matching of the extracted features agains a product database.
Task was to build a system that detects and recognizes grocaries.
Setup the collection and labelling of the data.
Designed and trained CNN detector that detects and classifies different features on products, (Name, EAN, Barcode)
Parse detections and pipe them into an OCR.
Implemented weighted matching of the extracted features agains a product database.
09/2022
-
09/2022
Intership - computer vision for enhancement of x-ray images
SMITH Detection
(Industrie und Maschinenbau, 250-500 Mitarbeiter)
Research project in cooperation with the Visual Inference Lab at the TU Darmstadt.
Task: Self-supervised Denoising and Demosaicing of X-Ray images.
Challenge: State-of-the-art computer vision methods are design for normally distributed RGB images. However, the data formats and most importantly the statistic distribution of X-Ray images are different.
This required an adaption of the methods before they could be applied to the task.
Experiments with designing a joint-network that simulatneously denoises and demosaics the raw sensor measurements resulted in a more performant model, but the state-of-the-art was not beaten.
Task: Self-supervised Denoising and Demosaicing of X-Ray images.
Challenge: State-of-the-art computer vision methods are design for normally distributed RGB images. However, the data formats and most importantly the statistic distribution of X-Ray images are different.
This required an adaption of the methods before they could be applied to the task.
Experiments with designing a joint-network that simulatneously denoises and demosaics the raw sensor measurements resulted in a more performant model, but the state-of-the-art was not beaten.
07/2020
-
10/2020
Software Developer ROS
TU Darmstadt
(Internet und Informationstechnologie, 1000-5000 Mitarbeiter)
Implementation of a ROS API for the Schunk EGH Gripper.
09/2019
-
09/2020
Data Science Working Student
Merck KGaA
(Pharma und Medizintechnik, >10.000 Mitarbeiter)
Implementation of the most recent advances in Deep Learning for research projects. (Graph Convolutional Networks)
Automatic plant monitoring, outlier detection and visualization dashboard.
Machine learned classification for automatic quality assurance.
Automatic plant monitoring, outlier detection and visualization dashboard.
Machine learned classification for automatic quality assurance.
Reisebereitschaft
Weltweit verfügbar
In Deutschland werktags alles möglich.
