Development of individual algorithms and software modules for data science, numerics and physics
Algorithmus Schmiede GmbH
New on Amplifa
About Algorithmus Schmiede
Algorithmus Schmiede GmbH is a contract research provider operating at the intersection of data science, numerics and physics. The company, founded in 2021 by Dr. Markus Dutschke and based in Saarland, develops individual algorithms and ready-made software modules (Python, C++, REST API on AWS) for industrial customers. The service portfolio covers complex optimization tasks in operations research, automated document processing, automated image data analysis and sensor technology, language processing, reinforcement learning, and machine-learning prototyping with experiments in production facilities. In a typical project, the company conceptualizes the task with the customer, builds a prototype, conducts experiments in the facilities, and ultimately delivers software modules for computing the target variable. The employees are PhD-qualified natural scientists; the focus is on reliability and prediction accuracy of the developed algorithms, since faulty forecasts in an industrial environment can lead to downtime or product recalls. Customers include Heidelberger Druckmaschinen, T-Systems, SONIVERSE and Dentsply Sirona.
- Processes
- Development of algorithms and software (data science, numerics, physics), Complex optimization tasks / operations research, Development of control algorithms (printing presses), Automated image data analysis / image processing, Automated document processing (categorization and extraction from PDF invoices), Object color determination from optical distance measurement (AI/sensor technology), Reinforcement learning (e.g. dice game Yahtzee), Language processing / language-processing systems
- Materials
- Complex Optimization / Operations Research, Machine Learning / AI, Image Processing / Sensor Technology, Language Processing, Reinforcement Learning, Software Development in Python and C++
- Founded
- 2021
- Employees
- 2
- Industries
- Industry & Manufacturing, Logistics and Transport, Retail, Airlines, Energy Supply, Healthcare
- 01
Project conceptualization with the customer
- 02
Prototype development
- 03
Experiments in production facilities
- 04
Algorithm development and benchmarking
- 05
Delivery of ready-made software modules (Python, C++, REST API on AWS)
Processes and Technology
Complex Optimization / Operations Research
Solving optimization problems that can no longer be solved with standard libraries, including Monte Carlo methods, simulated annealing and a modified Dijkstra algorithm.
Machine Learning / AI
Use of regressors, classifiers and neural networks, e.g. for document categorization, extraction and sensor data analysis.
Image Processing / Sensor Technology
Automated image data analysis, e.g. determining object color from optical distance measurement with 80% accuracy.
Language Processing
Development of language-processing systems as an own focus area alongside the industrial algorithms.
Reinforcement Learning
Development of learning systems, demonstrated e.g. by a superhumanly strong Yahtzee player.
Software Development in Python and C++
Delivery of ready-made software modules including testing frameworks and REST API on AWS (API Gateway, Lambda, RDS).
Projects and Services
Inside the production
Industries
Processes & technical limits
| Processes | Tolerance | Max. part | Weight | Lot size | Materials |
|---|---|---|---|---|---|
| Development of algorithms and software (data science, numerics, physics)Engineering | – | – | – | – | – |
| Complex optimization tasks / operations researchEngineering | – | – | – | – | – |
| Development of control algorithms (printing presses)Engineering | – | – | – | – | – |
| Automated image data analysis / image processingEngineering | – | – | – | – | – |
| Automated document processing (categorization and extraction from PDF invoices)Engineering | – | – | – | – | – |
| Object color determination from optical distance measurement (AI/sensor technology)Testing | Color determination with accuracy of 80% | – | – | – | – |
| Reinforcement learning (e.g. dice game Yahtzee)Engineering | – | – | – | – | – |
| Language processing / language-processing systemsEngineering | – | – | – | – | – |
| Construction of testing framework / software testing (Python)Engineering | – | – | – | – | – |
Additional services
Reference customers
History
- 2018
Dr. Markus Dutschke received his doctorate in theoretical physics from the University of Augsburg with an algorithm for more accurate evaluation of measurement data from ACAR experiments; afterwards working as an IT freelancer.
- 2021
Founding of Algorithmus Schmiede GmbH.
- 2022
Senior Python development for T-Systems International GmbH as part of the ECB project SPACE (Statistical Production and Compilation Environment), including construction of a virtual testing framework (January 2022 to February 2024).
- 2023
Lecture "AI for Companies" at the Junior Chamber Saarland (Wirtschaftsjunioren Saarland).
- 2024
Practical report on implementing custom AI solutions at AI Navigator 2024 in Nuremberg.