SPEAKER
Nandini Singhal
Senior Software Engineer, Datadog
Nandini Singhal is a Software Engineer at Datadog, where she operates Kafka streaming infrastructure at global scale, managing thousands of clusters processing petabytes of data daily. She has more than 8 years of experience building distributed systems in storage, with previous roles at Oracle Cloud Infrastructure and Microsoft. Her work spans distributed storage systems, concurrent programming, and performance optimization at scale.
Upcoming sessions
See complete program4 November
12:25 PM · 30 min
Operating Coding Agents at Scale
Rolling out coding agents across an engineering organization requires evidence about how they are used, whether they improve engineering work, and how each change affects quality, latency, and cost.
At Datadog, we combine gateway telemetry, normalized agent-session traces, task evaluations, and software-delivery outcomes. Gateway data captures model usage, errors, latency, and cost. Session traces show context and tool use, file changes, tests, and retries. Controlled evaluations compare models, harnesses, and configurations on the same tasks. Staged rollouts test those changes with engineers and monitor for regressions before broader availability.
Together, these signals guide rollout decisions, platform defaults, and cost controls without treating token volume, session count, or tool calls as measures of productivity.