Artificial Intelligence (AI) / Data Science / Machine Learning Consulting and Implementation: strategy and prototype development, feasibility assessment, individual coaching for executives. Koehn AI is a business consulting practice specializing in Artificial Intelligence. We support companies and non-profit organizations through strategy and prototype development, deployment and operations, feasibility assessments and individual coaching for executives. We build on experience from supporting a wide range of clients from Germany and beyond, amongst them start-ups, medium-sized enterprises, hidden champions and large corporations. In all we do, we strive to convey and condense even the most complex scenarios into accessible lines of reasoning.
Strategy Development
Artificial Intelligence or Adept Idiocy?
A solid AI strategy tremendously benefits organizations large and small.
The often very heterogeneous data landscapes of our clients hold an enormous value potential.
The approach to unlock this potential is given through Artificial Intelligence.
This results in considerable business value, or in process optimizations that enable a more productive work flow.
To this end, we work closely with executives and stakeholders.
We develop suitable metrics and KPI in order to measure the success and improve decision-making and business performance.
If you are interested in creating a solid AI strategy for your organization, or in challenging your existing one, send us a quick note via the web form or give us a call.
We can take the discussion from there.
Prototyping
You might have a concrete business case in mind and you might have a rough idea that AI could be a key ingredient,
but you are struggling to map out the concrete steps on how to tackle this problem.
Or you might want to rethink an existing prototype in light of the state of the art in AI research.
You might have heard about an intriguing new technique, and now you are curious to learn how it could help your business.
You might be faced with an extremely hard problem that none of your team members can solve.
We specialize in developing technical concepts and solutions for advanced machine-learning problems.
We help you synthesize the business and the technical aspects of your use case.
We are well-versed in all the major state-of-the-art frameworks and paradigms for (deep) machine learning, whether supervised, unsupervised or reinforcement learning.
Our proficiency in harnessing contemporary research for real-life use cases will help you to stay ahead of the competition.
We can work either on your systems, or use our powerful Nvidia workstation or cloud computing services.
Let us start the discussion. Send us a quick note via the web form or give us a call.
Deployment and Operations
Deploying machine-learning models takes about 25% of data scientists' time, while
75% of models never go beyond the experimental phase, resulting in a significant productivity bottleneck.
Automated deployment and retraining processes decrease the time to market and reduce operations cost.
This is the subject of machine-learning operations - or in short MLOps.
MLOps is an extrapolation of the DevOps approach to include the machine-learning modelling life cycle.
Models degrade over time because the input changes.
MLOps introduces continuous retraining, model monitoring and evaluation of performance.
It saves a version history not only of code, but also of data and models.
Versioning of data allows data scientists to keep track of where their data came from,
and versioning of models allows to efficiently keep track of the model quality in the development process.
We help you to streamline your deployment and operations process with model-training pipelines (TFX, MLflow, Pachyderm, Kubeflow),
registry, serving, monitoring and CI/CD orchestration.
Send us a quick note via the web form or give us a call to discuss your deployment and operations setup.
Feasibility Assessment
Artificial Intelligence is one of the most discussed topics these days when it comes to the development of new business models.
Numerous start-ups have undergone phases of considerable success and growth, spurring a surge of inspiration for others to follow on similar tracks.
For you as an investor, some of them will lead to huge returns, while most of them will never take off.
You might be thankful for some solid advice regarding the soundness of a business proposal and its technical viability.
A solid feasibility assessment of the new business case will give you a much better confidence level regarding your time and money investments.
Feel free to send us a quick message or give us a call to get the discussion underway.
Success Stories
We have successfully supported clients from sectors automotive, insurance, banking, telecommunications, food industry, engineering, retail and logistics. In this process, we have worked with our clients on topics such as automated feature generation from product photos, Bayesian analysis using TensorFLow Probability and PyMC (Bayesian regularization backprop using Levenberg-Marquardt optimization for an implementation of MacKay's method), automated correlation analysis using auto-sklearn, Ray and Dask, prediction of road-dynamics sensory measurement using LSTM and transformer networks for real-time applications, customer churn modelling using Pareto/NBD and neural networks, customer lifetime analysis, returns forecast modelling using a combination of RNN and Random Forest, conversion-delay predictions, dynamic marketing attribution modelling from the ground up using both Markov chains and Shapley values, time-series prediction and energy-trading strategy, materials assessment via image recognition, prediction of the inner clock from genetic samples, genetic algorithms in energy trading, KPI from tracking data in professional sports, NLP for customer-service workflow, image recognition for tumor detection, deployment of a TensorFlow model for predictive parking, automatic activity recognition in video data with LSTM and convolutional neural networks, dashboard development for traffic monitoring, engine test-bench analysis, statistical modelling for use cases in banking, scaling-plan for global roll-out of machine learning model, AI strategy development including definition of roles & processes and assessment of tools, execution of a range of workshops on AI strategy and use-case development, derivation of KPI from telemetry in free-floating car-sharing, supply-chain optimization, algorithmic trading strategies for stock market.