Security Engineer

Deepgram runs speech-to-text, text-to-speech and voice agents for Twilio, Cloudflare, Sierra and around 1,300 other organisations, and it runs them mostly on its own hardware. That is what makes this security job unusual.

Most of the infrastructure is bare metal in colocation datacenters, GPU-heavy, running containerized workloads on Docker, managed with Ansible and shipped through GitHub Actions, with AWS only for overflow and a few services. The posting draws the consequence itself: in a high-trust, fast-moving engineering culture, controls have to be delivered as code, be reproducible and produce their own audit evidence, because "anything that depends on someone remembering to do it manually will not survive". The surface area named is 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 behind SOC 2, PCI DSS and ISO 27001.

Read the company operating rhythm section before applying. Deepgram expects aggressive use of AI and agentic tooling, measures how effectively it is applied, and says outright that this is the wrong role if you want something prescriptive and nine to five. Remote, US.