Machine Learning Engineer

The Cake is seeking a Senior Machine Learning Research Engineer who operates at a research-grade level, primarily focused on inventing new machine learning methods and advancing the field through original research. The role involves conducting research on unstructured data, designing experiments, and translating validated research into production systems.

Responsibilities:

• Invent and evaluate novel machine learning methods
• Architectures, learning paradigms, optimization strategies, or representation frameworks
• Conduct first-principles research on unstructured and multimodal data
• Design rigorous experiments to validate novel ML ideas, including failure analysis and ablation studies
• Translate a subset of validated research into production-grade systems when appropriate
• Collaborate with senior researchers and engineers to develop defensible IP, including patents and internal research artifacts
• Author or co-author research papers, technical reports, and internal whitepapers
• Influence long-term R&D direction through original thinking and experimentation

Requirements:

• PhD in Machine Learning, Computer Science, Statistics, Applied Mathematics, or a closely related field
• 5+ years of experience post-PhD working on research-level or research-driven ML problems
• Demonstrated proof of original ML innovation, such as:
• Peer-reviewed publications with meaningful citations
• Patents or patent-pending ML methods
• Novel architectures, algorithms, or learning frameworks
• Deep experience with modern ML architectures (e.g., Transformers, generative models, representation learning)
• Strong proficiency in Python and at least one systems language (C++ or Java)
• Ability to take ideas from theory → experimentation → defensible implementation
• Have authored or co-authored highly cited machine learning research
• Have contributed to patents or protected intellectual property in ML
• Have experience in research labs or innovation-driven R&D teams (academic or industrial)
• Have created new datasets, benchmarks, or evaluation methodologies used by others
• Can clearly articulate what is novel about your work — and why it matters
• Have demonstrated long-term ownership of complex research efforts
• Thrive in ambiguous problem spaces where solutions must be invented, not assembled