ML Engineer III (Research Enablement)

About The Role
Accelerate machine learning research by building the engineering foundations, workflows, and shared tooling that researchers need to iterate at scale. Embedded within the AI team, you will work with researchers and engineers to turn evolving research needs into reliable, reusable capabilities.
Unlike a conventional MLOps role focused on deploying a single, well-defined model or feature, this role supports a broad and changing set of research workflows in which the models and ML pipelines are themselves the product. The ideal candidate has a strong platform engineering, MLOps, or DevOps/infrastructure background, understands the ML lifecycle, and is motivated to help researchers solve ambiguous problems faster.
Key Responsibilities
Partner with ML researchers across generative AI teams to identify bottlenecks and improve the speed, scale, and reliability of research iteration
Design, build, and maintain reusable research infrastructure, workflows, templates, interfaces, and automation for experimentation, training, evaluation, data processing, and model packaging
Enable reproducible experiments through consistent environments, dependency management, artifact and model versioning, configuration, observability, and CI/CD practices
Support scalable ML workloads involving large datasets, GPU clusters, distributed compute, and multiple interconnected models, services, and algorithmic components
Deliver pragmatic research-enablement capabilities for immediate needs while keeping them aligned with the architecture and roadmap of the central ML platform
Act as the technical bridge between researchers and the ML platform team: translate research pain points into clear platform requirements, validate new capabilities, and help research teams adopt the shared platform
Improve the path from research to product by making research outputs easier to reproduce, integrate, and test
Contribute to shared ML engineering standards and architecture across teams, and foster strong engineering practices through hands-on collaboration, technical guidance, and knowledge sharing
Evaluate and introduce technologies that materially improve research velocity, reliability, scalability, and cost efficiency
Requirements
Bachelor's degree in Computer Science, Computer Engineering, or a related field
4+ years of experience in platform engineering, DevOps, backend engineering, ML infrastructure, MLOps, or a closely related field
Proven experience building reusable infrastructure, tooling, or developer platforms that enable multiple engineers or researchers, not just a pipeline for one predefined model or feature
Strong proficiency in Python and Linux, including writing maintainable software, automation, and services
Hands-on experience with Docker, Kubernetes, CI/CD pipelines, cloud environments such as AWS, and Infrastructure as Code such as Terraform
Practical understanding of the end-to-end ML lifecycle, including data preparation, experimentation, training, evaluation, model and artifact management, packaging, deployment, and monitoring
Experience supporting compute-intensive or distributed workloads and diagnosing reliability, performance, resource, and cost bottlenecks
Working knowledge of modern ML frameworks such as PyTorch, with enough familiarity with model behavior and constraints to collaborate effectively with researchers (this is not a research scientist role)
Ability to work from ambiguous and evolving requirements, discover the underlying need, and turn it into simple, reusable engineering capabilities
Strong communication and collaboration skills across research, engineering, and platform teams
Core Competencies
Systemic thinking: see research workflows as an interconnected system spanning data, compute, experiments, models, evaluation, platforms, and production handoffs
Research enablement: empathy for researchers and a focus on removing friction without imposing solutions designed for narrower, fixed use cases
Problem-solving: technical excellence and curiosity when working through open-ended problems, incomplete requirements, and operational bottlenecks
Ownership: proactively find high-impact gaps, drive solutions to completion, and improve the environment beyond assigned tickets
Adaptability: comfort building for immediate needs while evolving solutions toward a longer-term shared platform
Communication and collaboration: translate effectively between researchers and platform engineers, build alignment, and share knowledge across teams
Nice to Have
Experience enabling research in ML domains such as language modeling, language translation, computer vision, multimodal or generative AI, robotics, and autonomous systems
Experience scaling GPU clusters for training, distributed computing, and large-scale data processing
Experience building researcher-facing ML platforms, self-service experimentation environments, and tooling used across many evolving research workflows
Experience helping research teams migrate to or adopt a shared ML platform
Knowledge of inference optimization techniques such as custom GPU kernels (valuable, but secondary to research enablement and ML infrastructure experience)
Additional Information
This position has no supervisory responsibilities

Work arrangement
Yes