Senior Security Engineer

LiteLLM•San Francisco, CA

2d•$220,000 - $260,000

About The Position

LiteLLM is seeking its first Security Engineer to secure its application, infrastructure, and software-delivery pipeline. This is a hands-on security generalist role with a strong emphasis on application security. The engineer will review and harden LiteLLM’s Python codebase, identify attack surfaces, collaborate with engineers on vulnerability remediation, and integrate secure development practices. The role also encompasses security work across IT, CI/CD, cloud infrastructure, and the software supply chain. The ideal candidate has experience in an early-stage security startup and a passion for security demonstrated through activities like CTFs or vulnerability research.

Requirements

A strong security generalist who can work across application security, IT, CI/CD, cloud infrastructure, and the software supply chain.
Deep application-security expertise, including manually auditing production Python code and working with engineers to remediate vulnerabilities.
Strong understanding of authentication, authorization, tenant isolation, SSRF, injection, deserialization, secrets management, and common web and API vulnerabilities.
Experience using and configuring tools such as Semgrep, Bandit, CodeQL, Burp Suite, and other SAST or DAST tooling.
Previous experience at an early-stage security startup during its 0→1 journey.
Experience securing containers, GitHub Actions, build pipelines, packages, and software dependencies.
Familiarity with SBOMs, Sigstore, Cosign, Snyk, Grype, Trivy, or equivalent tooling.
Strong knowledge of OAuth2, JWT, mTLS, IAM, and high-throughput API authentication.
Familiarity with prompt injection, LLM data exfiltration, tool abuse, and the OWASP Top 10 for LLM Applications.
Experience with incident response, CVSS scoring, vulnerability management, CVE triage, and coordinated disclosure.
Experience competing in CTFs during college or independently pursuing vulnerability research, reverse engineering, bug bounties, or security projects.
A genuine interest in security outside your day job—you regularly explore new attack techniques, build security projects, or contribute to the security community.
Bachelor’s or Master’s degree in Computer Science or a related field, or equivalent practical experience.

Responsibilities

Conduct deep security reviews of LiteLLM’s Python proxy, APIs, authentication systems, and enterprise features.
Identify and remediate vulnerabilities involving authentication, authorization, secrets, tenant isolation, injection, and data exposure.
Partner directly with engineers throughout design, implementation, code review, and release.
Build application-security tooling into the development lifecycle, including SAST, DAST, dependency scanning, and secrets detection.
Perform internal red teaming and adversarial testing against LiteLLM’s APIs, proxy, and LLM-specific attack surfaces.
Threat-model new products and architecture changes before they reach production.
Create secure coding guidelines and train engineers on common vulnerabilities and defensive practices.
Harden LiteLLM’s Docker images, PyPI packages, GitHub Actions workflows, and release infrastructure.
Detect dependency confusion, poisoned packages, compromised dependencies, exposed secrets, and unsafe build practices.
Implement SBOMs, signed builds, provenance checks, and reproducible-build practices.
Design secure-by-default configurations for cloud and self-hosted deployments, including authentication, IAM, secrets management, and key rotation.
Review cloud infrastructure, network boundaries, access controls, and production deployment patterns.
Strengthen employee identity, device, SaaS, and internal access controls.
Build monitoring and anomaly detection for suspicious API, model, authentication, and routing activity.
Lead security incident response, vulnerability assessment, remediation, post-mortems, and stakeholder communication.
Establish formal vulnerability intake, CVE triage, disclosure, and remediation processes.
Maintain threat models as LiteLLM’s product and architecture evolve.