Need to make the case to your manager?
The sessions below tackle these problems:
AI writes code faster than people can review it.
Min Chen, Principal Staff Engineer @LinkedIn, describes an AI code review system that handles tens of thousands of pull requests a week across about 10,000 repositories, with more than 60% of sampled suggestions accepted. Learn the architecture behind it, and how the team checks that suggestions reach the merged code.
Once agents are in production, tokens drive cost and latency.
Tejas Chopra, CEO @Headroom Labs, who worked on distributed caching infrastructure @Netflix, covers compression, caching, and retrieval for production LLM context. Take away three patterns with production numbers: learned prompt compression, prompt ordering that makes provider caches hit, and tiered storage for conversation history and tool outputs.
Running an LLM judge is the easy part of evals.
Fabio Flores, Software Engineer @DoorDash, leads foundation development for Agentic Ordering and shares the eval loop behind DoorDash’s ordering agent. Learn how to apply a single calibrated judge both offline and in production, and how the team checks that offline gains carry into production.
Agents don’t fit identity systems built for people.
Nancy Wang, CTO @1Password, keynotes on identity and authorization when software becomes a user. Learn how to narrow authority between subagents and grant access without exposing secrets in the context window.
An operations agent needs more than a good prompt.
Ronak Nathani, Principal Staff Software Engineer, Compute Infra @LinkedIn, explains the operations agent that helps run 500,000+ Kubernetes nodes. Take away server-side guardrails like rate limits, delete protection, and peer approval.
AI amplifies what a team already does well, and what it does badly.
Nathen Harvey, Lead of DORA @Google Cloud, speaks on building in an AI-first world. Assess your AI maturity beyond adoption metrics, using the DORA AI Capabilities Model.
Across the two days, you’ll return with patterns ready to adopt, costly mistakes to avoid, and a way to reduce the risk of what your team builds next. Use our convince your manager page for an email template and AI prompt to help build your request.
|
