Welcome to another Data Architecture Elevator podcast! Today's discussion is hosted by Paolo Platter supported by our experts Antonino Ingargiola and Irene Donato.

In this episode, we explore effective strategies for optimizing large language models (LLMs) for inference tasks with multimodal data like audio, text, images, and video.

We discuss the shift from online APIs to hosted models, choosing smaller, task-specific models, and leveraging fine-tuning, distillation, quantization, and tensor fusion techniques. We also highlight the role of specialized inference servers such as Triton and Dynamo, and how Kubernetes helps manage horizontal scaling.

Don't forget to follow us on LinkedIn! Enjoy!

Welcome to another episode of Data Architecture Elevator, hosted by Paolo Platter (https://www.linkedin.com/in/paoloplatter/) and featuring Antonino Ingargiola (https://www.linkedin.com/in/antonino-ingargiola/) and Antonio Murgia (https://www.linkedin.com/in/antonio-murgia/).

The conversation explores how Model Context Protocol (MCP), introduced by Anthropic, enables interoperability between LLM-based agentic applications and external systems through a standardized client-server model. In this setup, agents (LLM clients) discover and invoke tools exposed by MCP-compliant servers. This protocol decouples integration logic, allowing systems to interoperate without custom code. The discussion then applies MCP to data products in a data mesh architecture, where output ports (data access), observability ports (data quality and freshness), and control ports (operations) can be exposed as tools via MCP. This allows agents to autonomously query data, evaluate its quality, and even trigger operational commands like restarts or compliance actions. Enjoy it :)

Welcome to another episode of Data Architecture Elevator Podcast!

Today we'll tackle another hot topic: Data Privacy related to Large Language Models. Large language models (LLMs) have changed the way we interact with technology, enabling more natural conversations, automating tasks, and generating human-like responses. But as the saying goes, with great power comes great responsibility. The intersection of LLMs and privacy raises crucial questions: How do these models process and store user data? What safeguards exist to prevent unintended data retention? And what steps can businesses take to ensure compliance with data protection regulations? Join the discussion with our host ⁠Paolo Platter⁠ and our guests ⁠Antonio Murgia⁠, ⁠Antonino Ingargiola⁠ and⁠ Irene Donato⁠.

Don't forget to follow us on LinkedIn!

Enjoy!