Nico Jahn

Storm Reply is the Reply Group company focused on designing, implementing, and operating cloud-based solutions and services for large enterprises. As an AWS Premier Consulting Partner, it works across modern IaaS, PaaS, and SaaS environments. In this setting, I lead AI engineering and AWS/Azure cloud solution architecture work for enterprise clients. My role combines technical delivery with mentoring, knowledge sharing, and building reliable solutions across modern machine learning and cloud environments.

Data Reply is the Reply Group company for advanced analytics, AI-powered data services, modern data platforms, and production-ready machine learning solutions. Its work spans modern data architecture, big data engineering, DataOps, and MLOps for enterprise clients. In my senior consulting role, I mentored team members, ran workshops, and collaborated closely with stakeholders in fast-moving consulting settings. I delivered AI engineering, cloud, and container platform solutions for clients across legal, healthcare, consumer goods, manufacturing, and automotive. My work focused on Kubernetes, Terraform, GitHub Actions, and cloud-native delivery.

Data Reply helps organizations turn data into scalable analytics and AI solutions through modern platforms and engineering practices. The company combines data engineering, advanced analytics, and cloud delivery for enterprise use cases. As an IT Consultant, I worked in data engineering and machine learning for clients across healthcare, consumer goods, manufacturing, and automotive. I built solutions with Python, Java, C/C++, AWS, Azure, and Kubernetes in both cloud and edge-oriented environments.

Interautomation develops solutions for operating data logging, automatic passenger counting, real-time passenger information, and traffic control in public transport. The company focuses on improving rail passenger traffic and integrating those systems into real-world operations. In this research-driven industrial setting, I worked on machine learning and software development in Python, C, and C++. My work covered neural networks, inference workflows, and computer vision.

TU Berlin is one of Germany’s leading universities of technology and part of the Berlin University Alliance. It combines strong research, international exchange, and collaboration with society and industry. In the university IT environment, I provided support and administration across Windows, Linux, and macOS systems. My responsibilities included Active Directory, virtualization, software distribution, and day-to-day server operations.

Real-time load information in public transport is of high importance for both passengers and service providers. Neural algorithms have shown a high performance on various object counting tasks and play a continually growing methodological role in developing automated passenger counting systems. However, the publication of public-space video footage is often contradicted by legal and ethical considerations to protect the passengers’ privacy. This work proposes an end-to-end Long Short-Term Memory network with a problem-adapted cost function that learned to count boarding and alighting passengers on a publicly available, comprehensive dataset of approx.13,000 manually annotated low-resolution 3D LiDAR video recordings (depth information only) from the doorways of a regional train. These depth recordings do not allow the identification of single individuals. For each door opening phase, the trained models predict the correct passenger count (ranging from 0 to 67) in approx.96% of boarding and alighting, respectively. Repeated training with different training and validation sets confirms the independence of this result from a specific test set.

Automatic passenger counting (APC) in public transportation has been approached with various machine learning and artificial intelligence methods since its introduction in the 1970s. It is mainly used for revenue sharing, which (in Germany alone) is in the billions annually and supply planning, which is essential for services of general interest. While equivalence testing is becoming more popular than difference detection (Student’s t-test), the former is much more difficult to pass to ensure low user risk. On the other hand, recent developments in artificial intelligence have led to algorithms that promise much higher counting quality (lower bias). However, gradient-based methods (including Deep Learning) typically run into local optima. In this work, we explore and exploit various aspects of machine learning to increase the reliability, performance, and counting quality of the Neural APC 3D depth video-based LSTM neural network. We perform a grid search with several fundamental parameters: the selection and size of the training set, which is similar to cross-validation, and the initial network weights and randomness during the training process. Using this experiment, we show how aggregation techniques such as ensemble quantiles can reduce bias, and we give an idea of the overall spread of the results. We utilize the test success chance, a simulative metric based on the empirical distribution. We also employ a post-training Monte Carlo quantization approach and introduce cumulative summation to turn counting into a stationary method and allow unbounded counts. All in all, our numerous additions provide a major quality increase to the NAPC.