Senior Security Engineer

Deepgram is looking for a Senior Security Engineer to build and automate the technical controls behind our security program. Most of our infrastructure is bare metal in colocation datacenters — GPU-intensive, running containerized workloads on Docker, managed with Ansible, and shipped through CI/CD on GitHub Actions — with AWS used for overflow capacity and a small set of services. Our engineering culture is high-trust and fast-moving. That means controls have to be delivered as code, be reproducible, and produce their own audit evidence — anything that depends on someone remembering to do it manually will not survive.

You will own real surface area: host hardening and configuration management across the fleet, vulnerability and patch management, penetration testing, container security, detection and response, CI/CD and supply chain security, and the engineering work behind our SOC 2, PCI DSS, and ISO 27001 obligations. You will also be expected to use AI and agentic tooling aggressively to punch above your weight on a small team, and to help us secure the AI systems we build and the AI tools we adopt internally.

This is the hands-on technical seat. A GRC and Security Compliance Lead owns policy authorship, security questionnaires, and the auditor relationship; you own the controls themselves and the evidence they generate, and you give that role the technical answers it needs.

This role reports to the Director of Information Security.

Note: as reflected on our published compensation ranges, we are open on level. Strong mid-level and senior candidates are both in scope, and we will calibrate title, scope, and compensation to demonstrated experience.

• Build, deploy, and maintain security controls across our bare-metal fleet and AWS footprint — access management, network segmentation and firewalling, encryption and secrets management, logging and detection coverage, endpoint security, and vulnerability management.
• Own configuration management and host hardening as code with Ansible, so baselines are reproducible across hundreds of Linux hosts and drift is detectable rather than merely documented.
• Secure containerized workloads: Docker image build pipelines, registry scanning, runtime hardening and detection, and secrets management.
• Run vulnerability management end to end – discovery, prioritization by real exploitability rather than raw CVSS, driving remediation with engineering teams, and reporting on it.
• Own patch and update management as a program: the server fleet, the endpoint fleet, base images, OS and package updates, firmware where it matters, and dependency upgrades. Set and hold patch SLAs, track exceptions, and automate the routine cases so they do not depend on anyone remembering.
• Own the penetration testing lifecycle: scoping and vendor selection, scheduling, triaging findings into engineering work with real owners and dates, verifying fixes on retest, and producing the summary we can share with customers.
• Write detections, tune out noise, and take part in incident response. Improve log, telemetry, and security-agent coverage where we are blind.
• Run periodic log and access reviews as a standing control, and automate the collection so each cycle produces its own evidence rather than a scramble.
• Automate compliance evidence collection for SOC 2, PCI DSS, and ISO 27001, and support audit fieldwork on technical questions — the GRC lead owns the auditor relationship and the narrative; you own the systems that make the evidence true and repeatable.
• Harden GitHub Actions CI/CD and the software supply chain: dependency and secret scanning, action pinning, SBOM, artifact and image signing, least-privilege OIDC in place of long-lived credentials.
• Apply AI and agentic tooling to security work — triage, evidence gathering, code review, detection engineering — and help secure how the rest of the company uses AI.
• Provide the technical input behind customer-facing security work: questionnaire and architecture answers, penetration test summaries, and hardening guidance for customers running Deepgram self-hosted.
• Partner with research, ML infrastructure, and SRE teams to secure GPU training and inference systems. Threat-model them with the people who run them, and agree on controls before building.
• Own the roadmap for your workstreams: set priorities and milestones, keep stakeholders informed, and bring recurring risks and root causes to leadership.

• Solid experience in security engineering or infrastructure engineering with a security focus — enough that you have owned controls in production, not just recommended them.
• You’ve built a security roadmap with engineering or research leaders: you set milestones, got buy-in from the teams who own the systems, and delivered it. You can explain your trade-offs and the specific decisions you made, not just the frameworks you followed.
• Deep Linux. You should be comfortable hardening, debugging, and reasoning about a large fleet of physical Linux hosts — users and sudo, SSH, systemd, kernel and package updates, host firewalls, and what a host-based agent is actually doing.
• Ansible at fleet scale is required. You automate first and document second.
• Strong scripting. Python and/or Go, plus real shell competence.
• Production experience securing Docker containers and the pipelines that build them.
• Hands-on experience securing GitHub and GitHub Actions: branch protection, CODEOWNERS, workflow permissions, secrets, and third-party action risk.
• Experience running vulnerability and patch management as an ongoing program — including the unglamorous part, which is getting other teams to actually ship the fix.
• Experience managing third-party penetration tests from the inside: scoping them well, and turning a report into closed engineering work rather than a PDF in a folder.
• Hands-on involvement in at least one formal audit — ISO 27001, PCI DSS, or SOC 2 — as the engineer producing and defending evidence, not only as a reader of the report.
• Comfortable using AI coding and agentic tools in daily work, and clear-eyed about their risks: prompt injection, secret leakage, and supply chain exposure.
• Pragmatic. Controls that block shipping without a proportionate risk reduction get routed around, and we would rather you name the trade-off than pretend it does not exist.
• Hands-on experience securing self-hosted infrastructure: physical Linux servers in datacenters or colocation, including network segmentation and out-of-band management (IPMI/BMC). Cloud-only experience is not sufficient.
• Experience securing GPU or ML infrastructure, such as GPU servers, training or inference clusters, or the platforms research teams run on.

• Detection engineering or SIEM/HIDS ownership at scale (for example Wazuh, OSSEC, Elastic).
• Secrets management with HashiCorp Vault.
• AWS security (IAM, SCPs, VPC) and Terraform.
• Kubernetes security.
• Offensive security background: penetration testing or red teaming.
• PCI DSS work inside a cardholder data environment.
• Ownership of a secure SDLC program.
• Certifications such as OSCP, GIAC, CISSP, or AWS Security Specialty.

Notice: We're aware of individuals impersonating Deepgram recruiters. All legitimate Deepgram recruiting communication comes from an @deepgram.com http://deepgram.com email address. If you've received a message claiming to be Deepgram, please forward it to careers@deepgram.com.

Deepgram is a research company that productizes as fast as possible. Four Operating Principles shape how we think, decide, and work together:

• Compress before you communicate. You think it through first, then lead with the bottom line.
• Resist entropy. You question what already exists before adding more process or tools.
• Solve the category. You fix recurring problems at the source, not one case at a time.
• Work at the frontier. You go after problems no one has solved yet.

We look for people who already work this way, and these principles shape how we grow and evaluate our team.

We build AI that millions of people talk to, and we run the company the same way. AI is how we centralize work here, and is core to how we operate.

• Build, not just use. Design your own workflows and agents, and use AI to remove whole categories of work instead of doing the same work faster.
• Own the output. AI can speed up the work. It doesn’t make anyone less accountable for it. You understand, verify, and stand behind everything you share, including what an agent produces for you.
• Spend the time you get back at the frontier. The capacity AI creates goes toward problems no one has solved yet.
• Keep pace. Models and tools change every month, and so will your work. You adopt what’s new quickly and drop what no longer earns its place.

We judge AI by what it helps you deliver and whether you stand behind the result, not by how often you use it. In interviews, expect to walk us through a workflow you built and where you drew the line between AI and your own judgment.

Company overview

Deepgram is bringing the world’s most natural interface, voice, to the technology we use every day. Our speech models and real-time infrastructure power AI systems that keep up with the way people actually speak. More than 5,000 organizations run conversational voice AI on Deepgram in production, including Twilio, Cloudflare, Sierra, Decagon, Vapi, Cresta, Granola, and Jack in the Box.

The Deepgram Voice Agent Platform brings together speech-to-text, text-to-speech and speech-to-speech, with voice-native foundation models available through cloud APIs or self-hosted and on-premises deployments. Built for accuracy, low latency and cost efficiency, Deepgram makes high-quality voice experiences practical to build and run at scale. More than 650,000 developers have used Deepgram to process more than 200,000 years of audio. Every day, more of the world’s conversations with AI run on our models.

About Deepgram

Deepgram builds foundational voice AI – speech-to-text, text-to-speech, and voice agent infrastructure — used by enterprises and developers to build production voice applications at scale. Our models are trained and served on our own infrastructure, and we offer hosted, dedicated single-tenant, and self-hosted deployment options so customers can meet their own data residency and retention requirements. Learn more at deepgram.com http://deepgram.com.