InfoQ Homepage Large language models Content on InfoQ
-
Fine Tuning the Enterprise: Reinforcement Learning in Practice
Will Hang and Wenjie Zi explain how OpenAI's Agent RFT (Reinforcement Fine-Tuning) optimizes reasoning models end-to-end, enabling autonomous agents to learn efficient tool use and cut latencies.
/presentations/rft-openai-model/en/mediumimage/WenjieZiWillHang-medium-1782220624463.jpg)
-
Graph RAG: Building Smarter Retrieval Workflows with Knowledge Graphs
Cassie Shum explains how to move beyond traditional RAG limits using GraphRAG to inject enterprise logic and complex multi-hop reasoning into scalable knowledge graphs within a data warehouse.
/presentations/graph-rag-llm/en/mediumimage/CassieShum-medium-1782291352027.jpeg)
-
Rules for Understanding Language Models
Naomi Saphra explains why language models act like populations rather than individuals. She discusses how data diversity drives generalization, why LLMs mirror users and how tokenizers alter behavior.
/presentations/5-principles-llm-behavior/en/mediumimage/naomi-saphra-medium-1781688751052.jpg)
-
From Hype to Strong Foundations: What the Rise, Fall and Resurgence of Agents Can Teach Us about Outlasting the Cycle
Aditya Kumarakrishnan discusses "Agents: The Missing Manual," sharing four historically grounded ideas to build modular, durable, and hyper-tenant AI agent architectures.
/presentations/llm-compound-ai-systems/en/mediumimage/medium-1781082183297.jpg)
-
Building Evals for AI Adoption: from Principles to Practice
Mallika Rao explains how evaluation debt silently triggers regressions in distributed AI systems. She shares a five-layer evaluation stack to align metrics directly with long-term user trust.
/presentations/eval-ai-adoption/en/mediumimage/medium-1779185675202.jpeg)
-
Accelerating LLM-Driven Developer Productivity at Zoox
Amit Navindgi explains how Zoox built Cortex, an internal AI platform that streamlines the developer lifecycle by moving beyond the hype to deliver secure, agentic workflows and real-world impact.
/presentations/ai-software-development/en/mediumimage/medium-1778065503665.jpg)
-
Dynamic Moments: Weaving LLMs into Deep Personalization at DoorDash
Sudeep Das and Pradeep Muthukrishnan discuss how DoorDash combines LLMs with deep learning to move from "classic" collaborative filtering to "hyper-personalization" in real-time commerce.
/presentations/llm-personalization/en/mediumimage/Sudeep-Das-Pradeep-Muthukrishnan-medium-1776173227456.jpg)
-
Building Embedding Models for Large-Scale Real-World Applications
Sahil Dua explains the architecture and training of embedding models. He shares practical tips for distilling large models and scaling RAG applications for real-time production environments.
/presentations/llm-large-scale-applications/en/mediumimage/sahil-dua-medium-1769590214923.jpeg)
-
Ecologies and Economics of Language AI in Practice
Jade Abbott explains how to build sustainable AI using "Little LMs." She discusses environmental impacts, linguistic justice, and technical optimizations like quantization and model distillation.
/presentations/llm-ai-africa/en/mediumimage/jade-abbott-medium-1766064709735.jpeg)
-
Why Observability Matters (More!) with AI Applications
Sally O'Malley shares how to build an AI observability stack with open-source tools (Prometheus, Grafana, OpenTelemetry, Tempo, vLLM/Llama Stack). Learn to track performance, quality and cost signals.
/presentations/observability-llm/en/mediumimage/SallyOMalley-medium-1757669042150.jpg)
-
Deploy MultiModal RAG Systems with vLLM
Stephen Batifol explains the core concepts of multimodal RAG systems, vector search indexes (HNSW, IVF), and embedding model selection. He details vLLM and Pixtral for optimized inference..
/presentations/rag-vllm/en/mediumimage/stephen-batifol-medium-1758016495689.jpg)
-
Chatting with Your Knowledge Graph
Jonathan Lowe discusses how to enable an LLM to chat with a structured graph database. He explains the process of using semantic search and knowledge graphs to answer natural language questions.
/presentations/llm-graph-database/en/mediumimage/jonathan-lowe-medium-1756898332000.jpg)