Innovation Group
Analytics proof of concept, then production
Most new analytics ideas die in a slide deck or a notebook. We test yours on your own data and tools, drop the dead ends early, and turn the ones that work into software people use.
How we work
First make it work. Then make it usable.
We are strict about correctness before we spend money on screens and deployment. Then we design with the people who will use it.
Start on your stack
Python, KNIME, Power BI, Databricks or Siemens software. We prototype in what your team already runs.
Work with the users
We sit with operators, planners and managers to learn what they need to decide, and keep checking with them as we build.
Decide early
A proof of concept ends with a clear call: advance it, or drop it before it eats the budget.
Rebuild the winners
What works gets rebuilt as a production app, with the plumbing that makes it last: screens, workflows, reports and deployment.
The work
Kinds of problems we take on
For the past decade our Innovation Group has worked as a standing partner to large manufacturers, closing scores of projects every "semester". A sample:
- Long-range capacity planning: multi-year facility and equipment plans.
- Staffing: workforce models and scheduling tools for managers.
- Warehouse sizing: DC capacity and layout.
- Quality: sampling-rate tools, anomaly detection on process data, vision-based checks.
- Material and energy use: scrap and usage tracking with reporting leaders can read.
We mix whatever fits: simulation (including discrete rate, which we started), optimization, statistics and machine learning.
The team
Operations research, economics and code
The group brings master's and PhD-level training in analytics, economics, operations research and machine learning. Its technical lead, Mathias Brandewinder, served on the board of the F# Software Foundation and wrote Machine Learning Projects for .NET Developers.
We use F# to explore data and to build the core of a model, because the same code can grow into a production app. That keeps the path from idea to software short.
Where ideas went
- 2000s
VINLogic
A 6-hour logistics model rebuilt as a 20-minute app. It convinced us the future was apps, not just models.
- 2020s
Edge apps
Line-side tools that coach operators in real time. See Edge & IIoT.
- 2024
A spinoff
Ideas that became products moved to ChiAha, so SDI could stay focused on client work.
From a client
"The SDI team went above and beyond the project scope, proactively brainstorming ways to make our project better."
J Adam Traina, Director of Operations Research, Symbotic. Read the warehouse case.
Tell us your problem
What would happen if you changed it?
Describe the decision in front of you. We'll tell you whether a model can answer it, and what it would take.
- A straight answer on whether simulation is the right tool
- Which tool fits, even if it isn't one we use every day
- A rough scope and timeline, before any commitment