Simon Geisler
Verified Expert in Engineering
AI Researcher and Developer
Munich, Bavaria, Germany
Toptal member since March 19, 2020
Simon is an AI and data science professional with more than five years of experience in data science, data engineering, software engineering, and system architecture design in the public cloud (products scaled to 10+ million users). In Germany and Silicon Valley, he worked on spatio-temporal IoT, natural language processing, and computer vision projects. Besides his strong affinity to modeling algorithms, Simon confidently develops production-grade microservices, Spark, and Kafka Streams apps.
Portfolio
Experience
- Computer Vision - 7 years
- Artificial Intelligence (AI) - 6 years
- Python - 5 years
- Apache Spark - 4 years
- Deep Learning - 3 years
- Azure - 3 years
- Docker - 3 years
- PyTorch - 2 years
Work Experience
Research Associate
Technical University of Munich
- Had more than 10 publications at tier-one machine learning conferences, including NeurIPS, ICML, and ICLR.
- Was a main contributor to research projects with Google and Fraunhofer.
- Handled teaching graduate-level machine learning courses.
Research Intern
Google DeepMind
- Extended the popular transformer architecture to directed graphs (main application: source code).
- Published a research paper at ICML 2023 and the Deep Learning on Graphs workshop at AAAI 2023.
- Awarded a "contributed talk" at the Deep Learning on Graphs workshop at AAAI 2023.
Senior System Architect
Robert Bosch GmbH
- Reworked the system design for the road condition service (automated driving) on Azure using the latest services.
- Designed and implemented data analytics and machine learning pipelines using Spark, Airflow, and Azure Data Explorer.
- Improved scalability of the wrong-way driver warning in regions of 100+ million users.
- Set up the development process from writing tickets until deployment.
Deep Learning Intern
Bosch Center for Artificial Intelligence (BCAI)
- Exceeded state of the art performance with multiple approaches for the task of hazard detection (one of the long tail tasks for automated driving).
- Engineered custom deep learning architecture incorporating for real-time (embedded) object detection using stereoscopic vision.
- Prepared cutting-edge research publications currently in review.
Data Scientist | AI Engineer
Robert Bosch GmbH
- Developed a detection algorithm for tracking vehicles based on GPS and Open Street Map based on particle filtering.
- Designed and implemented a Kafka Streams architecture in Scala for out of order handling and scaling the detection algorithm (improved scalability to 100+ million users).
- Implemented CI/CD pipelines in GitLab for building the project, executed tests of different granularities and license/dependency/security scanning, and dockerized deployment on Kubernetes.
- Implemented various microservices with Node.js, namely for persisting data in the Azure Data Explorer (proprietary big data telemetry storage).
- Developed various features and made production-ready on an Angular-based front end.
Research and Development Intern
Robert Bosch GmbH
- Supported several projects for Advanced Driver Assistance Systems (ADAS) focusing on digital maps and electronic horizon.
- Organized and supervised user experience surveys and test group studies.
- Developed an Android app for rapid prototyping of HMIs for a highway assistant that communicates via a Bluetooth CAN interface.
- Developed an Android app (end-to-end) for hazard spots on the road, including the cloud service.
Project History
Cloud-based Wrong-way Driver Warning
I contributed to this project in various ways. Most importantly, I developed a particle filter-based detection algorithm for probabilistically spotting this needle-in-a-haystack phenomenon. Then as the technical lead, I led a team of two other data scientists on spatiotemporal data. I developed various big data pipelines with Apache Spark to crunch gigabytes of graph data. To scale the service, I led the development of a Kafka Streams architecture implemented in Scala.
In the DevOps process, it was also essential to develop further artifacts such as CI/CD pipelines for the Kubernetes-based architecture.
Stereo Vision Hazard Detection
I delivered PyTorch and Tensorflow implementation of various architectures. I could exceed the current state of the art in a relative performance gain of more than 10%.
Education
PhD in Computer Science
Technical University of Munich - Munich, Germany
Master of Science Degree in Data Science
Albstadt-Sigmaringen University and University of Mannheim - Albstadt, Mannheim, and Tübingen in Germany
Bachelor of Engineering Degree in Mechatronics and Micro Systems Engineering
Heilbronn University - Heilbronn, Germany
Certifications
Software Architecture
oose Innovative Informatik eG
Software Quality
Robert Bosch GmbH (internal training)
Deep Learning Specialization
deeplearning.ai
Developer Training for Spark and Hadoop
Cloudera
Statistical Learning (with Distinction)
Stanford University
Basic Certificate in Project Management (GPM)
GPM Deutsche Gesellschaft für Projektmanagement (German Society for Project Management)
Skills
Libraries/APIs
Matplotlib, Scikit-learn, PyTorch, Pandas, TensorFlow, OpenCV, NumPy, Web API, Azure API Management, Spark ML, Node.js, AMQP, Shapely, JAX
Tools
Jupyter, GitLab CI/CD, Kafka Streams, Git, MATLAB, Apache Storm, Torchvision, Azure Kubernetes Service (AKS), Terraform, Azure Machine Learning, GitLab, Splunk, Grafana, Apache Airflow, Spark SQL, Azure DevOps Services, Jira, Bitbucket
Languages
Python 3, Python, SQL, OWL, Java, Scala, JavaScript, C++, XML, YAML
Frameworks
Spark, Apache Spark, Hadoop, Yarn, Angular
Paradigms
Lambda Architecture, Extreme Programming, Spatial Databases, Automation, Continuous Integration (CI), Continuous Delivery (CD), Web Architecture, Anomaly Detection, DevOps, Scrum, Normalizing Flows
Platforms
Apache Kafka, Jupyter Notebook, Azure, NVIDIA CUDA, Docker, Android, Kubernetes, Databricks, Azure Event Hubs
Storage
Databases, PostgreSQL, MongoDB, Redis, Data Pipelines, Data Integration, Google Cloud, SQL Architecture, Azure Blobs, Azure SQL Databases, Azure Cache, PostGIS, JSON
Industry Expertise
Automotive, Project Management
Other
Agile Deployment, Machine Learning, Deep Learning, Data Mining, Data Analysis, Data Analytics, Artificial Intelligence (AI), Artificial Neural Networks (ANN), Data Science, Statistical Learning, Predictive Modeling, Torch, Big Data, Big Data Architecture, Algorithms, Image Classification, Semantic Segmentation, Object Detection, Stereoscopic Video, Microsoft Azure, Computer Vision, Recommendation Systems, Natural Language Processing (NLP), Autonomous Robots, Semantic Web, Time Series, Spatial Analysis, Localization, Simultaneous Localization & Mapping (SLAM), Quantitative Modeling, Geospatial Data, Geospatial Analytics, OpenStreetMap, Data Engineering, Generative Pre-trained Transformers (GPT), Web Mining, Programming, Knowledge Bases, Ontologies, Scientific Computing, Engineering, Signal Processing, Digital Signal Processing, Statistics, Bayesian Statistics, Numba, Statistical Methods, Naive Bayes, Decision Trees, Decision Analysis, Decision Modeling, Statistical Modeling, Multivariate Statistical Modeling, Natural Language Understanding (NLU), Software Design, Software QA, Software Architecture, System Design, High Code Quality, Quality Management, Distributed Systems, Data Architecture, Supervised Learning, Classification, Text Classification, Sentiment Analysis, Unsupervised Learning, Regression, Embedded Hardware, Cloud Services, Azure Data Factory (ADF), Azure Virtual Networks, Azure Virtual Machines, Monitoring, Internet of Things (IoT), HTTP, Clustering, Active Learning, Robotics, Autonomous Navigation, Automotive Controller Area Network (CAN), GPU Computing, Data-driven Decision-making, Controls, Deep Reinforcement Learning, Graphs, Graph Neural Networks (GNNs), Transformers, Adversarial Robustness, Government Performance Management (GPM), Reinforcement Learning
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring