Detection & Response Engineer

The Role

You'll build the detections and automation that protect our cloud, SaaS applications, and user workstations. You'll be the team's Datadog expert, owning how security telemetry gets into Datadog Cloud SIEM, how it's filtered, and the detections that run on it. You'll also automate the repetitive work around alerts and support triage and investigations. As we bring automation and AI into security operations, you'll help build and ship the new capabilities.

What You'll Own

• Detection engineering: write, tune, and retire Cloud SIEM detections, and map coverage to MITRE ATT&CK to close gaps.
• Logging pipeline: onboard sources and build Datadog log pipelines that parse, normalize, and enrich security telemetry.
• Detection-as-code: manage detections in version control with testing and peer review.
• Automation: build workflows for alert enrichment, triage, containment, and notification.
• AI in the SOC: build AI-assisted triage and alert summarization, with a human in the loop.
• Measurement: track coverage, false-positive rate, and automation rate to decide what to fix next.
• Filtering and cost: manage indexes, exclusion filters, and retention tiers to cut noise and cost without losing coverage.
• Security operations: support triage, investigations, and incident response, including on-call.

What You Bring

• 3+ years in security operations, detection engineering, or security engineering.
• You write your own scripts and build detection logic as code. Infrastructure as code and CI/CD experience is a plus.
• Deep hands-on Datadog experience, especially Cloud SIEM detection rules and log pipelines.
• Experience with automation orchestration tooling.
• Working knowledge of MITRE ATT&CK. Threat hunting or purple teaming experience is a plus.
• Experience with security telemetry from cloud platforms, SaaS applications, and EDR.
• You can take a project from prototype to production with little direction.

The Ideal Candidate

The ideal candidate measures success by real threats caught and noise removed, not by the number of detections written. Automating a problem is their first move, and they would rather ship a working detection this week and tune it than wait for a perfect rule. They know Datadog well enough to test every parsing and detection change against real logs before calling it done. They use AI to move faster but verify its output, and when they find a coverage gap or a noisy log source, they fix it instead of filing a ticket.

#LI-LB2