Radha Basu, the Founder and CEO of iMerit, spoke with Analytics Insights in an exclusive interaction about how the DataOps framework can be used to tackle edge cases in AI. She also elaborates on how the company is efficiently investing on its clients as well as its employees for a better future..
Here are 3 key takeaways from the podcast:
- MLOps = ML ModelOps + ML DataOps
High-quality data, precision, continuous feedback, technology propelled by expert intelligence, and the fusion of ML modelOps and ML DataOps in an infinite loop is what MLOps is about. Today, the focus in terms of getting AI into production is evolving from a model-centric approach to a data centric approach.
- Handling edge cases in AI
To tackle edge cases, iMerit has designed a product called the iMerit Edge Case where technology and knowledge management are tied together to find the edge cases in every dataset that automated models breakdown, and work in the best way to handle them and similar cases going forward.
With iMerit Edge Case, clients can easily identify the edge cases, trigger a collaborative workflow to redefine the requirements, document and share the edge cases, view edge case-related analytics and insights, and create a repository of edge cases for future projects.
- Privacy and security are critical
Clients trust iMerit with one of their most valuable assets – their data. iMerit has developed best practices across the organization to keep huge data volumes safe throughout the labeling process. A dedicated Information Security Manager is responsible for maintaining security, and she works in tandem with key internal and external stakeholders. iMerit centers have set up CI/CD infrastructure, are SOC 2 Type II certified, ISO 27001:2013 certified, ISO 9001:2015 certified, and GDPR and HIPAA compliant.
Listen to the Analytics Insight podcast here.
