AI at scale: The Snowflake Cortex way

At YourStory DevSparks 2025, where developers from across India gathered to build smarter, faster, and for the future, the spotlight was on real, usable innovation, not hype. And when Harish Chintakunta, Senior Partner Solution Engineer at Snowflake, took the stage for his workshop, the message was simple: you don’t need a PhD in machine learning to build GenAI-powered applications anymore.

The session, titled ‘Unlock the True Potential of Your Data with Snowflake Cortex’, delivered exactly what it promised: an end-to-end walkthrough of how developers can embed powerful GenAI features directly into their applications and workflows using tools they already work with, inside the Snowflake platform.

Chintakunta began by setting the stage for Cortex’s role in the evolving data landscape. “The biggest hurdle has been marrying AI and live data in a seamless, scalable way,” he explained.

Traditional AI models require heavy infrastructure and constant data movement, slowing down innovation. Cortex breaks this mould by being fully serverless and tightly integrated with Snowflake’s cloud data platform. This means AI can operate on live, governed data directly, no data lakes or pipelines needed. “It’s a paradigm shift,” Chintakunta emphasised, “giving enterprises the agility to innovate faster with their data.”

What is Snowflake Cortex?

Cortex is Snowflake’s fully managed, serverless AI platform, built to seamlessly integrate generative AI capabilities within the company’s data cloud ecosystem. By eliminating the complexities of infrastructure management, Cortex empowers organisations to build AI applications faster and securely at scale, leveraging the power of Snowflake’s cloud data platform.

Cortex simplifies AI adoption with out-of-the-box support for leading large language models (LLMs), scalable GPU-accelerated compute, and developer-friendly tools. This allows teams to embed AI directly into their data workflows and business applications with ease.

Core capabilities: Powering enterprise innovation

Diving deeper, Chintakunta showcased Cortex’s core strengths: cloud-agnostic architecture, automatic scaling, and built-in data governance. “Security is non-negotiable,” he remarked.

Cortex’s native integration with Snowflake ensures that AI workflows respect enterprise governance policies. Additionally, the drag-and-drop studio makes AI accessible to both data scientists and business users alike, democratising AI development. “By lowering the barriers to entry, we enable teams to prototype and deploy AI solutions rapidly,” he noted, highlighting how this capability accelerates time to value.

Inside Cortex AI: Intelligent models tailored for business

Cortex runs on a serverless architecture that automatically scales CPU and GPU compute, and supports AI workloads across AWS, Azure, and Google Cloud. It offers infrastructure for deploying models using Nvidia GPUs and container runtimes, alongside a model registry for easy deployment and management.

“You can bring your own models or use ours, whatever fits your use case,” Chintakunta said. The platform also includes a suite of enterprise-optimised generative models tailored for semantic search, summarisation, and automated insights. “Our AI models grasp the subtle context within enterprise datasets, delivering precise, actionable intelligence instead of generic responses,” he added.

This ability to interpret complex business data transforms raw inputs into ready-to-use insights. To support developers and data scientists, Cortex offers a no-code Studio, native notebooks, and Streamlit integration for building interactive applications, all backed by Snowflake’s governance and security framework.

Cortex Analyst: Democratising data access

Recognising that not all users are data experts, Chintakunta introduced Cortex Analyst, a no-code interface that allows users to interact with Snowflake data using natural language. “With Analyst, you can ask questions like ‘What were Q4 revenues in South India?’ and get an answer without writing SQL,” he explained. “Imagine a product manager simply asking, ‘What were the sales trends last quarter?’ and instantly getting detailed answers.”

Cortex Analyst applies semantic modelling to interpret domain-specific vocabulary, translating everyday questions into accurate SQL queries. This eliminates the dependency on technical teams, delivering instant, self-service insights and fostering a more inclusive, data-driven culture across organisations.

Hands-on lab: Bringing Cortex to life

The workshop concluded with an interactive hands-on lab, where participants created Snowflake accounts, loaded datasets, and configured Cortex tools such as Search and Analyst. Guided by Chintakunta, attendees built interactive web apps using Streamlit that supported natural language queries, enabling semantic searches and lightweight AI workflows.

“Within an hour, people were querying datasets using plain English and seeing the power of AI embedded into apps,” Chintakunta said. “Seeing is believing. The lab demonstrated Cortex’s real-world practicality, proving it’s not just a concept but a scalable solution ready to address today’s enterprise challenges.”

This exercise showcased how quickly and easily AI capabilities can be deployed by non-specialists, emphasising Cortex’s intuitive design and accessibility.

Why this matters

The Snowflake Cortex workshop at DevSparks 2025 showed how organisations can integrate AI directly into data systems without needing extra infrastructure or complex model management. With tools like Cortex Analyst and Studio, Snowflake makes AI development accessible to more teams, not just data scientists.

As Chintakunta concluded, “Cortex helps businesses activate their data with AI, securely, at scale, and without friction.”