DevSecOps Engineer

Job description

About The Role We handle some of the most sensitive data around: patient outcomes, clinical trial records, and the multimodal datasets behind our foundation model, all inside a company that runs on agents with broad, standing access to internal systems. That combination raises the stakes on security in a way most companies never have to think about. The threat landscape is moving fast too. Capable AI models have lowered the skill floor for finding and exploiting vulnerabilities, so the volume and sophistication of attacks aimed at companies like ours keeps climbing. We need someone who takes that seriously, builds for it before it's a problem, and doesn't treat security as something you check off at the end of a project. We're hiring a DevSecOps engineer to own security architecture and practice across our AI and data infrastructure: the pipelines, agent tooling, and internal systems that power our BD, clinical development, computational biology, and lab teams. This isn't a compliance-only role. You'll design the guardrails, then get in the code and build them yourself. What You'll Build Threat models and security architecture for a system built around large numbers of autonomous AI agents with access to internal tools, data, and MCP-style servers. Guardrails, sandboxing, and permissioning for how agents talk to systems and each other, so they can act without excessive standing privileges. Secure CI/CD pipelines, infrastructure-as-code review, and automated security testing (SAST/DAST, dependency and secret scanning) built into how we ship, not added on after. Monitoring, detection, and incident response tooling tuned for AI-native infrastructure, including anomalous agent behavior, prompt injection attempts, and data exfiltration paths. Data governance and access controls for a governed warehouse and knowledge graph holding patient-level and clinical trial data, aligned with HIPAA and relevant regulatory frameworks. Cloud security posture across our GCP environment: IAM, network segmentation, encryption, audit logging. Who You Are You've owned security for a production system end to end, from architecture through detection and response, somewhere a breach would be a real problem, not a hypothetical one. You think like an attacker and build like an engineer: you don't just write policy, you ship the code and tooling that make the secure path the easy path. You've kept up with how much easier AI has made attacks to pull off, and you turn that into concrete engineering decisions instead of vague concern. You move quickly, take initiative, and want to be judged on systems that hold up under real pressure, not paperwork that satisfies an audit. Must-haves Roughly 6+ years in security engineering, DevSecOps, or something close, with real ownership of production security architecture. Fluent in Python and/or TypeScript. You can read and write production code, not just configure tools. Deep experience with cloud security (GCP preferred, AWS or Azure fine too), IAM design, network security, and secrets management. Hands-on experience securing CI/CD pipelines and infrastructure-as-code (Terraform or similar). Practical experience building or operating detection and incident response systems, not just reading dashboards someone else set up. Comfortable working as an individual contributor with engineers. Once the security foundation is solid, you'll spend real time as a working engineer on the broader platform. Bonus Experience securing agentic AI systems, LLM-powered applications, or MCP-style tooling, including prompt injection defense and agent permissioning. Experience in regulated environments handling patient or clinical trial data (HIPAA, GxP, or similar). Certifications like OSCP, CISSP, or GCP/AWS security certs are a nice signal but not a requirement. Location This is a hybrid role, requiring 4 days per week onsite, in our NYC Headquarters. Drug development shouldn’t be guesswork, not when patients are waiting. Pathos is building a next-generation biotech with AI at the core. Not as a feature, but as the operating system for how medicines get developed. We believe most drugs don't fail because the science was wrong. They fail because they were tested in the wrong patients, with the wrong assumptions, in trials that couldn't answer the real question: who benefits, and why? Pathos exists to change that. We're building the largest foundation model in oncology and pairing it with proprietary AI systems, deep oncology expertise, and 200+ petabytes of multimodal data linked to patient outcomes, so we can make development decisions with more precision, much earlier. This is not theoretical. We’re well-capitalized and have the leadership to build a generational company. We invest in and advance our own clinical-stage programs, using our AI platform to sharpen trial design, patient selection and biomarker strategy. So therapies reach the patients most likely to benefit, sooner. How We Build Pathos does not operate like a traditional biotech. There is no middle management. There are no layers of approval. The company is designed, from the ground up, around small teams of 2–4 subject matter experts who each command hundreds of AI agents to do the work that used to require dozens of people. Everyone builds. Everyone ships. Every function at Pathos — from clinical execution to asset selection to the foundation model itself — runs on this model. Our product velocity delivers meaningful outcomes in hours instead of weeks. This is not a future aspiration. It is how we operate today. The people who thrive here are operators: deep experts who can specify what needs to happen, orchestrate AI agents to execute at scale, and make high-judgment calls that compound over time. If you have spent your career building and shipping AI systems at scale, this is the environment where that experience becomes a superpower.