ML projects have multiple stakeholders with varying levels of involvement and responsibilities. Early involvement and effective collaboration with stakeholders is essential for developing the right solution, managing expectations, and ultimately for a successful ML implementation.
As early as possible, define your project's stakeholders, the expected deliverables, and the preferred communication methods.
Be sure to include them in your list of stakeholders, as well as any other teams who need to approve aspects of your ML solution.
Deliverables
Each stakeholder might expect different deliverables at each phase of the project. Here's a list of common deliverables.
Design doc. Before you write a line of code, you'll most likely create a design doc that explains the problem, the proposed solution, the potential approaches, and possible risks. Typically, the design doc functions as a way to receive feedback and address questions and concerns from the project's stakeholders.
Experimental results. You must communicate the outcomes from the experimentation phase. You'll typically include the following:
- The record of your experiments with their hyperparameters and metrics.
- The training stack and saved versions of your model at certain checkpoints.
Production-ready implementation. A full pipeline for training and serving your model is the key deliverable. At this phase, create documentation for future engineers that explain modeling decisions, deployment and monitoring specifics, and data peculiarities.
You should align early with your stakeholders on their expectations for each phase of the project.
Keep in mind
In some cases, stakeholders might not understand the complexities and challenges of ML. This can make getting projects prioritized and executed difficult. For example, some stakeholders might assume that ML is similar to traditional software engineering practices with deterministic outcomes. They might not understand why the project's progress is stalled or why a project's milestones are non-linear.
To manage stakeholder expectations, it's critical to be clear about the complexities, timeframes, and deliverables at each stage of your project.