Expert insight
Building a modern Data & AI stack with Snowflake: the KPC masterclass
Architecture, governance and automation: the KPC method for building a robust data platform.
From the evolution of the Data Warehouse to the integration of generative AI, discover why this platform is establishing itself as an essential foundation for businesses.
In this new episode of Robin Conquet's DataGen podcast, Mickael Kuentz, Data & AI Director at KPC, reveals the keys to building a high-performance Data and AI architecture using Snowflake.
Snowflake, unleashing data to better exploit it
Snowflake was born from a simple promise: separate compute from storage to overcome the limitations of traditional platforms. The result: ultra-fast processing, even on very large volumes of data. As Mickael Kuentz points out, " Snowflake has unleashed data " by making the smooth querying of billions of rows possible.
From data warehouse to complete platform
Initially designed as a Data Warehouse, Snowflake quickly expanded its scope. Today, the platform covers all workloads of the modern data stack: ingestion, transformation, visualization, machine learning, generative AI, and even application development.
With tools like Streamlit and Snowpark Container Services, Snowflake now allows you to create applications and integrate AI models at the very heart of the data.
Advantages and limitations to keep in mind
Main advantages:
Simplicity and speed: streamlined implementation and administration.
Scalability: suitable for both startups and large corporations.
Functional coverage: a complete range of use cases.
Points to watch: The main risk remains vendor lock-in, even though Snowflake is increasing its interoperability efforts (Apache Iceberg, Microsoft partnerships).
Generative AI: the underlying trend
Snowflake is making generative AI a major strategic focus. AI now assists governance (e.g., Copilot Horizon) and accelerates the implementation of business use cases. In parallel, Snowflake is becoming a secure foundation for running GenAI models (OpenAI, Mistral, Anthropic, Meta).
Tips for a successful Snowflake project
Validate organizational prerequisites: maturity, clear use case, internal sponsorship.
Build a tailored stack : choose the right third-party tools according to your needs.
Democratize data: make the platform accessible to business profiles.
Anticipate AI : integrate AI into your data strategy starting now.
Why listen to this podcast?
This podcast is a true masterclass for data leaders, CTOs, and IT decision-makers.
Mickael Kuentz shares not only his technical expertise but also practical and strategic advice based on several years of project experience.
You will discover how Snowflake allows you to combine performance, simplicity, and innovation, while remaining attentive to governance and sovereignty challenges.
If you want to understand concretely how to build a high-performance Data & AI stack, this episode will give you valuable keys and inspiring feedback.
