"Agents often perform well during the experimentation phase, but things can quickly go wrong when they have to operate inside a real company's services, data, and processes."
"Serving LLMs at Scale: The Hidden KV Cache Advantage covers KV cache as the hidden lever behind inference cost and performance, with direct impact on GPU utilization, throughput, and 'Time to First Token'."
QCon AI Boston 2026 will take place on June 1-2 at Boston University, addressing engineering challenges in AI production. The program emphasizes the gap between impressive AI demos and systems that perform reliably under production conditions. Key topics include context engineering for AI agents, inference economics, and infrastructure. Sessions will explore how to build organizational context layers for AI, manage inference costs, and optimize performance in large-scale enterprise environments, ensuring AI systems are effective and sustainable in real-world applications.
Read at InfoQ
Unable to calculate read time
Collection
[
|
...
]