Machine Learning Consulting Process
1. Business Context
Any problem demands context from the business. A solution for one industry may not be applicable to another, nor is every business the same. Establishing shared context helps get the project off to the right start.
2. Stakeholder Inclusion
More than just identification, we find projects are optimal when key stakeholders are including in the project. With “skin in the game”, stakeholders are far more likely to collaborate to produce a better overall solution.
3. Problem Definition
A key phase where business problems are defined and prioritized. It is worth spending time to get this right, as subsequent effort could be ineffective and wasted.
Data scientists and machine learning engineers are already experts in the work that they do, so interviewing them to learn what they do, how they work, and what the models they have produced do, is an important stage in the process.
5. Model Review
Reviewing models is sometimes the main task in an ML consulting project. Here we collaborate with you to evaluate your current models. We look at performance, toil, risk, and many other important factors.
6. Solution Review
Given knowledge of the problem and the processes and models involved we’re now in a position to ask whether the problem is being solved. If not, we will isolate why not, and provide actionable insight into how to fix it.
7. Problem Solution
If necessary, we provide details as to what needs to happen in order to fix the issues you are having with this particular problem.
8. Strategic Solution
It is common that we find larger, strategic-level issues that should be addressed for optimal ML usage. In this phase we present these solutions back to key stakeholders. In many projects we then move into an implementation phase. Other times we step back and perform another iteration to fix another problem.