Audit your current AI initiatives to establish your company's AI maturity level.
Which AI Maturity Level Are You At?
1. Playing with AI
PoC projects are proving that AI can be applied to business problems
Models in development but not in production
Data science team small or outsourced with manual collaboration
2. Unsafely deploying
Deploying models into production
Initial wins with models but not yet scaling, and team working inefficiently and treading on each others' toes
Missing or incomplete audit trail and minimal production monitoring
3. Operationalized AI
Full reproducibility & provenance, audits not a problem
Model health monitoring throughout lifecycle
Collaboration in place to allow team to scale safely
How Production–Ready is your AI process?
We will also evaluate your current capabilities against the DevOps for ML Manifesto and provide a report. Here are some sample excerpts:
Low Scores on the Certification
Poor reproducibility: ad-hoc work done with occasional version control for code, no version control for data or models
Poor accountability: difficult to track back from a model to how it was created or to see on what basis it made decisions
Manual collaboration: ad-hoc collaboration done in meetings, by sending each other notebooks, or folder structures
Manual lifecycle: manual process to deploy model to production, no monitoring of model once in production
High Scores on the Certification
Perfect reproducibility: every run of every data transformation or model training can be reproduced, even years later
Strong accountability: every human decision that went into creating a model can be forensically reproduced
Seamless collaboration: the status of every experiment is seamlessly synchronized to the team's ML knowledge base
Continuous lifecycle: automated deployment of models to production with statistical monitoring of model drift
The Dotscience DevOps for ML Certification is a one-off cost of $10,000. This includes 5 hours of 1:1 consultation with our expert team and a report packed full of insights and recommendations on how to improve your AI process.