Tuana Çelik

LLMs are great at reasoning, planning, and orchestration, but prediction and classification are a fundamentally different task. Ask an AI agent to forecast a sensor reading, or score a loan application, and you're asking a language model to do something it was never built for: producing a calibrated, grounded number rather than a plausible-sounding one.

Most developers start with basic RAG implementations, but production AI systems demand sophisticated multi-step reasoning, as well as additional tooling. This talk demonstrates building production-ready agent workflows using the newly released LlamaIndex Workflows 1.0.

Some of the latest big evolutionary steps in generative AI has been models that support function calling and “agentic” capabilities. This is provides generative models with “tools” that allow them to go beyond generating outputs for simple queries, and start planning the best way to solve...

NLP applications and LLMs have incredible potential to transform the way we live, work, and play. But NLP applications, especially at production scale, can be confusing.

Retrieval-Augmented Generation (RAG) is an incredibly powerful technique (a hack, if you will) to make use of powerful LLMs while making sure they produce accurate information based on your own data. However, model providers and open-source frameworks cannot know exactly what you need for every application you might need to build.