Principal Engineer, AI Security

Lila Sciences•Cambridge, MA

About The Position

As a Principal Security Engineer focused on AI Security, you will define and drive the technical strategy for securing how AI is used across Lila's enterprise. You will operate as a senior individual contributor, partnering with IT and business teams to ensure safe and compliant adoption of AI tools and platforms. While Lila builds AI-powered systems, this role is primarily focused on securing the use of third-party and internally deployed AI tools across the enterprise — ensuring sensitive data, intellectual property, and scientific workflows are protected as AI becomes deeply embedded in how work gets done.

Requirements

8+ years of experience in information security, with strong expertise in enterprise, cloud, or application security.
Hands-on experience designing and implementing security controls in enterprise environments.
Familiarity with AI/ML systems and how modern AI tools (LLMs, copilots, APIs) are used in practice.
Experience with cloud platforms (AWS/GCP), SaaS security, and zero trust architectures.
Experience with data protection technologies (e.g., DLP, data classification, access controls).
Practical experience with threat modeling, red teaming, or adversarial testing.
Strong communication and influence skills across technical and non-technical stakeholders.

Nice To Haves

Experience securing enterprise use of LLMs, copilots, or generative AI platforms.
Familiarity with AI gateways, prompt filtering, or model interaction controls.
Experience evaluating or securing third-party AI vendors and APIs.
Background in regulated environments (biotech, healthcare, defense, or government).
Experience with browser security, endpoint controls, or SaaS security platforms.
Knowledge of privacy-enhancing technologies or confidential computing.
Contributions to AI/ML security research or community.

Responsibilities

Enterprise AI Security Strategy — Define and implement security controls and guardrails for the use of AI tools (e.g., LLM APIs, SaaS AI platforms, and internal AI services) across the organization.
AI Gateway & Agentic Gateway Security — Design and implement AI gateway controls to manage and monitor access to external and internal AI systems. Secure agentic workflows by enforcing identity, authorization, tool-use constraints, and policy controls for autonomous or semi-autonomous agents.
AI Red Teaming & Adversarial Testing — Conduct red teaming and adversarial testing focused on enterprise AI usage, including prompt injection, data exfiltration, jailbreaks, and abuse of connected tools and plugins.
Data Protection for AI Usage — Develop and enforce controls to prevent sensitive data leakage through AI systems, including input/output filtering, data classification, tokenization, and secure handling of prompts, embeddings, and outputs.
Multi-Layer AI Security (Network, Endpoint, Data) — Integrate AI security into existing enterprise security layers: network visibility and control over AI service access, API traffic inspection, and zero trust enforcement; endpoint security for developer machines, research environments, browsers, and plugins; data layer controls ensuring proper handling of sensitive data when interacting with AI systems.
AI Threat Modeling (Enterprise Context) — Develop threat models focused on enterprise AI usage, including risks such as data leakage, prompt injection, model misuse, supply chain risks from AI vendors, and unauthorized agent actions.
Vendor & Platform Security — Assess and guide secure adoption of third-party AI vendors and platforms, including evaluating data handling practices, model behavior, and integration risks.
Incident Response for AI Usage — Define and support response approaches for AI-related incidents, such as sensitive data exposure, policy violations, or misuse of AI tools.
Cross-Functional Technical Leadership — Partner with Legal, Compliance, IT, and Engineering to align AI usage with regulatory requirements, data governance policies, and responsible AI practices.
Security Enablement — Contribute to internal guidance and education on safe AI usage, including secure prompting, data handling, and appropriate use of AI tools.
Security Tooling & Implementation — Evaluate and implement tooling for AI security, including AI gateways, DLP integrations, monitoring solutions, and policy enforcement mechanisms.