The client is seeking a VP of Research, Machine Learning to set and evolve the research direction for its core intelligence, including context representation, memory, reasoning, planning, and orchestration. The role involves deciding when to design new model architectures versus adapting or leveraging frontier open-source or commercial models, and defining evaluation frameworks that measure real-world usefulness, robustness, safety, and long-term behavior rather than benchmark performance alone. The VP will own alignment, safety, and guardrail strategy as first-class product concerns and guide exploration of frontier techniques such as retrieval-augmented training, mixture-of-experts, distillation, multi-agent orchestration, and multimodal systems.
The VP of Research will shape early product intelligence direction in close partnership with product and application engineering, ensuring the technical bar for research rigor, judgment, and taste is consistently applied across the organization. This is a high-ownership role that emphasizes operating with strong product impact and making informed, potentially irreversible or high-impact decisions with incomplete information. The position is expected to focus on building systems that work in real-world conditions, with particular attention to model behavior, failure modes, and long-horizon trade-offs.
Qualified candidates will have deep experience building or evolving real machine learning systems used in production, along with strong technical judgment about model behavior, failure modes, and long-term trade-offs. The client is looking for a builder mindset with an obsession for evaluation, correctness, and how systems behave over time, coupled with a founder-like mentality and high ownership. Experience with GPU-based training and inference systems and comfort working with Python and modern ML frameworks are expected, while candidates whose primary focus is primarily publishing, incremental benchmarking, or managing a large research organization may not be a fit.