Google machine learning course
Overview
CLIENT: Google engEDU (internal training for Google engineers)
ROLE: Learning experience designer
AUDIENCE: Software engineers at Google who planned to work on a machine learning project
TOOLS: Markdown, HTML/CSS, Google Draw, Google’s developer documentation site
COURSE LINK: https://developers.google.com/machine-learning/data-prep/
Opportunity
Google's engEDU team (engineering education) wanted to convert its Machine Learning Bootcamp for employees into a self-study experience available to the public. Data Preparation is a core course in this series.
Process and solution
I analyzed the learner, the context, and current training. I then presented my recommendations to the stakeholders in a design doc that included learning objectives, assessments, instructional strategy, and an evaluation plan. I re-structured the workshop content, fleshed out examples with SME input, wrote practice questions, and designed visuals.
I also facilitated in-person pilot testing of the beta version. Participants used a guide I created to give verbal feedback after their review and also filled out an online survey. Feedback from pilot testers was very positive, even comparing more favorably to the ML crash course (a pre-req).
Impact
The beta feedback guide I created was so useful in helping us get constructive, specific feedback from target learners that the wider team began to use it for other course pilot tests.
Course analytics as of January 24, 2019 (includes internal usage data since July 2018):
-
Total visits per unique user: 26,246
-
Average time spent on course: 2 hrs and 38 mins (out of 4 hours total)
-
# of unique interactions: 3,510
We planned to measure more than the above completion metrics in a Phase 2 that would include a hybrid version with expert mentors.