Senior Research Engineer - Security Foundation Models (Zürich)

📋 Description

• • Join Lakera.Ai as a Senior Research Engineer and be at the forefront of AI security, pushing the boundaries of scalable AI security solutions. We are not just a research lab; we are actively engineering the next generation of security foundation models designed for immediate, large-scale impact. As a foundational member of our team, you will play a pivotal role in shaping our strategic approach, influencing critical decisions, and constructing robust systems that ensure the security of AI applications and agentic systems at an unprecedented scale.
• • Our mission is to build the next generation of security foundation models, venturing into uncharted territories of AI security. To achieve this ambitious goal, we are seeking world-class research engineers who possess the expertise to design and scale the essential infrastructure that underpins our innovative work. This role transcends mere support for research; it is about actively driving it forward. You will operate at the cutting edge of distributed training, meticulously optimizing large-scale Large Language Models (LLMs) from their fundamental architecture, and engineering sophisticated systems that propel the frontier of AI security.
• • Your responsibilities will encompass the entire spectrum of scaling training across numerous GPUs, optimizing inference processes for maximum efficiency, and being an integral part of building the foundational infrastructure that enables the secure deployment of AI at scale. This is an opportunity to contribute to a company that is defining the future of AI safety and security.
• • About Lakera: Lakera is dedicated to ensuring that AI systems behave as intended, aligning with human goals and values. We are moving towards a future where AI agents will manage significant aspects of our businesses and personal lives. At Lakera, we are not merely envisioning this future; we are actively constructing the security foundation upon which it will be built. We empower security teams and developers, enabling their organizations to confidently adopt AI technologies and unlock the next era of intelligent computing.
• • We collaborate with Fortune 500 companies, innovative startups, and leading foundation model providers, offering them protection for themselves and their end-users against adversarial misalignment. Furthermore, we are the creators of Gandalf, recognized as the world's most popular AI security game, demonstrating our commitment to engaging and educating the community.
• • Lakera maintains offices in San Francisco and Zurich. Our work environment is characterized by rapid progress and intense dedication. We operate as a unified team, yet we empower every individual to assume substantial ownership and accountability for their contributions. Transparency is a core value, upheld at all organizational levels, and we are unwavering in our commitment to consistently elevate our standards across all endeavors. We actively promote diversity of thought, firmly believing that it is the catalyst for achieving the best possible outcomes.
• • Example Projects: You will have the opportunity to contribute to projects such as scaling the training of large-scale security foundation models across vast parameter spaces. This includes designing and optimizing distributed training pipelines for highly efficient LLM post-training processes. You will implement reinforcement learning-based post-training methodologies for LLMs at scale, and develop sophisticated adversarial training pipelines engineered to harden AI systems against a wide array of real-world threats. A significant part of your role will involve engineering robust Machine Learning (ML) infrastructure to provide unwavering support for high-performance AI security research.
• • About You: You possess a passion for rapid development and a knack for building. You excel in designing and scaling AI systems, with a proven ability to translate models into production-ready solutions. You are invigorated by the challenges of real-world AI security problems and view engineering as the primary means to address them effectively at scale. You thrive in a dynamic, impact-driven environment where you can take ownership of the entire ML stack and push the boundaries of what's possible. We are particularly interested in candidates who demonstrate at least one of the following:
• • A proven track record of advancing research through dedicated engineering efforts, whether by implementing novel ML techniques, successfully scaling complex experiments, or significantly optimizing training and inference for state-of-the-art models.
• • Extensive experience architecting and optimizing large-scale distributed systems, spanning domains such as AI, cloud infrastructure, networking, or other similarly demanding environments.
• • A demonstrable history of scaling impactful research initiatives, ensuring that innovative ideas can transition seamlessly from conceptualization to large-scale deployment.
• • Hands-on experience scaling LLM training across multiple GPUs or TPUs, with a focus on optimizing for both efficiency and peak performance.
• • Proven experience in building resilient and robust ML pipelines that effectively support large-scale training and inference operations in production settings.
• • While an advanced degree in Machine Learning, AI, or a closely related field is advantageous, it is not a strict requirement. Our primary focus is on demonstrable, real-world impact and practical expertise over formal credentials. If you are eager to push the limits of AI security and contribute to building at scale, we encourage you to connect with us.
• • Stay Connected: To learn more about the exciting developments at Lakera, follow us on LinkedIn, Twitter, and Instagram. Join our community on Momentum, the dedicated Slack channel for AI Safety and Security professionals. To manage your information in our recruitment database, please reach out to [email protected].