Jobs at LawZero

Architect, implement, scale, and maintain petabyte-scale data processing platforms for frontier AI models. Build reliable, observable, cost-efficient pipelines with heterogeneous and GPU-accelerated workloads, including failure recovery and restartability. Ensure datasets and intermediate outputs are versioned, reproducible, traceable, and documented with datasheets. Partner with AI researchers to integrate datasets into training pipelines and guide sustainable technology decisions.

Develop and optimize distributed machine learning training and inference systems for very large models. Collaborate with research scientists, diagnose performance bottlenecks, profile and debug experiment code, build orchestration tools, optimize compute resources, and establish best practices for large-scale ML workflows.

Translate advanced AI research into scientifically accurate content for policymakers, funders, partners, media, and the public. Develop blogs, research summaries, white papers, newsletters, briefings, and stakeholder communications; support grant applications and technical research funding materials; and assist with conference, outreach, and partnership communications. The role requires strong AI and machine learning fluency, exceptional English writing, professional French proficiency, and at least five years of relevant experience.

Lead LawZero’s talent acquisition function by managing and developing recruiters, designing global sourcing and relocation strategies, overseeing employer branding and candidate experience, and personally recruiting for specialized technical roles. Build scalable hiring processes, ATS workflows, reporting dashboards, and competitive intelligence programs while partnering with Talent, Culture, and Marketing teams. The role requires international technical recruiting expertise, people leadership, and fluency in English and French.

Design and implement novel machine learning models for advanced AI safety research. Collaborate with mathematicians and research scientists, adapt and fine-tune frontier models, create experimental protocols and evaluation frameworks, analyze results, and optimize model performance with ML engineers. The role requires deep learning research experience, expertise with major ML libraries and distributed computing, strong Python development skills, and effective interdisciplinary communication.

Leads machine learning research and engineering teams focused on AI safety and the Scientist AI approach. Responsibilities include managing and mentoring researchers and engineers, hiring and onboarding technical talent, directing complex cross-functional projects, optimizing reproducible research workflows, advancing foundation-model research, and translating scientific concepts into large-scale engineering execution.

Develop and optimize machine learning research workflows for large-scale model training and inference. Collaborate with research scientists to implement novel models and simulated environments, build tools that improve computing efficiency, and establish best practices for ML development. The role requires advanced ML expertise, deep learning research experience, high-performance computing, cloud platforms, containerization, and data infrastructure familiarity.

Develop and evaluate probabilistic inference methods, including amortized inference for high-dimensional distributions and probabilistic graphical models. Translate theoretical research into high-quality Python implementations, design evaluation strategies, analyze experiments, and collaborate with mathematicians. The role also requires communicating complex findings and contributing to research on safe-by-design, non-agentic AI systems.

Develop and scale web-scale data pipelines that transform raw text into high-quality machine learning datasets. Responsibilities include deduplication, model-based scoring, filtering, toxicity and PII removal, metadata extraction, dataset versioning, provenance tracking, compliance, leakage detection, quality monitoring, and researcher tooling. The role also involves acquiring and evaluating text corpora, collaborating with research, engineering, legal, and governance teams, and developing novel data transformations for AI training.

Own and operate the platform supporting AI research workloads, including Kubernetes clusters, CI/CD pipelines, cloud infrastructure, internal services, observability, and security controls. Manage infrastructure as code, connect Kubernetes with HPC and Slurm environments, establish platform standards, document systems, automate recurring work, and participate in incident response. The role requires close collaboration with researchers while maintaining secure, reliable, and usable infrastructure.

Supports LawZero’s executive leadership through high-volume email and calendar management, cross-entity coordination, international travel planning, expense tracking, and administrative process improvement. The role acts as a liaison between LawZero and Mila, triages executive requests, coordinates meetings and special projects, and develops tools and protocols to improve operational efficiency. Strong discretion, organization, diplomacy, independent judgment, and fluent French and English are required.

Conduct foundational AI safety research with Yoshua Bengio and the LawZero team. Develop theoretical frameworks, safe AI architectures, and probabilistic safety guarantees under finite data and compute constraints. Analyze existing research, formalize safety requirements, guide empirical experiments, publish scientific work, and communicate complex ideas to technical and non-technical audiences. The role focuses on building Scientist AI systems that are capable, transparent, and safe by design.

Leads responsible AI and data governance practices, including data provenance, privacy, licensing, AI governance controls, research integrity, dual-use risk, bias and fairness testing, and internal AI tool oversight. Partners with legal, data, research, product, evaluation, and IT teams to establish policies, procedures, documentation, and compliance frameworks.