CXOToday has engaged in an exclusive interview with Kavita Viswanath, General Manager, JFrog India, on JFrog acquiring Qwak AI
Can you explain how the acquisition of Qwak AI will enhance JFrog’s existing platform and services?
JFrog already offers the most advanced ML Model Registry with JFrog Artifactory and additional AI/ML security and governance capabilities with JFrog Xray and Advanced Security. Through this acquisition, JFrog’s solution will be expanding to deliver advanced MLOps capabilities to organizations, enabling the building, deployment, management, and monitoring of AI workflows from GenAI and LLMs to classic ML models, all within a unified platform. As part of the JFrog Platform, Qwak’s technology will combine a straightforward and hassle-free user experience for bringing models to production with the level of trust and provenance enterprises expect to have when delivering AI-powered applications. This combination leverages Qwak’s advanced model training and serving capabilities with model storage management, versioning and scanning provided by JFrog.
How will this acquisition impact your current customers? What benefits can they expect to see?
Adding Qwak’s technology to the JFrog Platform is expected to free data scientists from worrying about infrastructure and application delivery, fueling JFrog’s MLOps commitments and powering customers’ rapidly growing AI imperatives. The union of JFrog and Qwak will provide customers with:
- One Platform for DevSecOps & MLSecOps, offering a holistic ML software supply chain from traditional models to LLMs and GenAI.
- Fast and straightforward model serving into production with simplified model development and deployment and serving processes, improving AI initiatives
- Model training and monitoring with out-of-the-box dataset management and feature store support.
- Manage models as a package allowing you to version, manage, and secure models the same way you do any other software package with DevSecOps best practices.
- Ensure provenance and security of AI naturally in the development workflows.
- Pull from a governed, secure source of truth that marries ML models with the other building blocks such as containers and Python packages.
- Trace models back to their source for easy recall, retraining, and redeployment if something goes wrong with production models.
What are the key features of Qwak AI’s MLOps solution that made it an attractive acquisition for JFrog?
As noted earlier, when it comes to AI, bringing model services to production is difficult and often fails. Data scientists and ML engineers currently use tools that are mostly disconnected from standard DevOps and Security processes within companies, delaying release timeframes and eroding trust. Qwak’s offering covers the full ML development and deployment of models to production including DataOps, ModelOps, and RuntimeOps, which will help pave the way for companies to secure the delivery of responsible AI models, simply and predictably – which is very compelling for both developers and C-suite leaders.
Can you discuss more about JFrog’s “model as a package” approach and how it will be integrated with Qwak’s technology?
JFrog’s “model as a package” approach enables fully traceable, reproducible, and immutable models. Treating models as packages is essential for them to run the same way between training and production environments, while allowing organizations to rollback models if performance issues arise. Today’s market demands a single platform experience across the software supply chain to accelerate development processes and treat machine learning (ML) models (and their metadata) like any other software component. ML models must be stored, built, traced, versioned, signed, secured and efficiently delivered across systems to deliver AI at scale. The holistic, JFrog + Qwak solution aims to eliminate the need for separate tools, separate compliance efforts and will offer full traceability in a single solution.
How will this acquisition help in accelerating the creation and delivery of AI-powered applications?
The potential of machine learning and Generative AI is reshaping the expectation and prospects for what applications can do, and in turn, what organizations must deliver to their customers to remain competitive. However, while AI model development has gotten more advanced in recent years, bringing model services to production is still very challenging and often fails because there are no well-established engineering or security best practices for developing and releasing ML models to production. Many organizations are missing visibility, governance, traceability, complete integration, and trust.
AI development must be brought into a fully automatable AI/ML pipeline to ensure it is part of the organization’s trusted, unified, secure software supply chain. With the acquisition, JFrog aims to deliver a unified and scalable solution for DevOps, Security, and MLOps stakeholders. This advanced, industry-leading MLOps functionality is designed to free data scientists and developers from infrastructure concerns, accelerating the creation and delivery of AI-powered applications.
How does the acquisition align with JFrog’s overall strategy and vision for the future?
JFrog’s Liquid Software vision demands a completely secure, seamless, automated pipeline to deliver software updates from code to any device. ML Models are rapidly becoming a standard component of all software releases, but organizations struggle to bring model services to production in a secure, trusted way. The incorporation of Qwak’s technology into the JFrog Platform will provide an end-to-end MLSecOps solution that brings the development and management of ML Models in line with an organization’s existing software development lifecycle and ensure that JFrog is the only solution needed to manage the components of modern Software Supply Chains.
The post JFrog Acquires Qwak AI; Enhancing AI Workflow Management appeared first on CXOToday.com.