Context Is All That Matters: MCP and the Dawn of Agentic Intelligence
MCP, AI Infrastructure, Agentic AI, LLMOps, OSS AI
Every generation of AI has been limited by context. Models could reason, but not act. They could predict, but not connect.
The Model Context Protocol (MCP) changes that forever. MCP is the connective tissue between models, tools, and data—the open protocol that lets any LLM see, know, and do. It turns isolated chatbots into context-aware agents that can sample, chain tools, browse resources, and collaborate with humans and systems alike.
In this talk, we’ll rip open the MCP spec and show how:
1) Clients and servers negotiate capabilities like context streaming, sampling, and prompting. 2) Tool chaining lets models plan and execute multi-step workflows autonomously.
3) Resources give every model live, composable access to your world—files, APIs, databases, and more.
4) Sampling orchestration unleashes multi-turn, multi-model creativity inside a single session.
You’ll see how MCP transforms LLMs into ecosystems: interoperable, composable, and ready to act in the real world. Forget prompt-engineering—this is protocol-engineering for intelligence.
Gabriela de Queiroz is the Founder of f02 labs, a consulting studio helping startups accelerate visibility, product adoption, and developer engagement through AI Strategy and Developer Relations as a Service.
Previously Director of AI at Microsoft, she advised hundreds of startups on building and scaling AI products, and before that led global AI strategy and open-source initiatives at IBM. She has also worked with startups across different stages, contributed to open-source projects, and serves as an advisor helping founders define their AI strategy and go-to-market approach.
In addition to her industry work, Gabriela teaches on platforms such as Maven, Coursera, and edX, sharing practical lessons at the intersection of AI, product development, and leadership.
She is also the founder of R-Ladies and AI Inclusive, global communities empowering more than 200,000 members in 65+ countries.
Shub Argha
Head of Solutions Engineering @ Arcade.dev
Shub builds and scales infrastructure that developers actually want to use. At ngrok, he worked with engineering teams across industries to design secure, production-ready ingress patterns everything from OAuth and mTLS to Kubernetes-native gateways and IoT networking. He helped companies move fast without compromising on security, and often served as the technical lead in high-stakes proof-of-concepts and large-scale platform rollouts.
At Arcade.dev, Shub focuses on the intersection of LLMs and real-world tools. He helps teams go beyond chat by wiring up AI agents to take secure, authenticated actions whether that’s calling APIs, integrating with third-party systems, or coordinating multi-step workflows.
His background spans DevOps, embedded AI, and developer tools, and he’s especially interested in making agentic systems feel reliable, observable, and easy to adopt.