The people building the platform that thousands of developers rely on every day.
MiniMax is led by four executives whose combined experience spans academic AI research, big-tech engineering, GPU infrastructure, and published machine learning science.
The leadership team works from the San Francisco headquarters and remote offices across North America and Europe. They operate on a simple principle: decisions are made by the people closest to the problem. The CEO sets vision and culture. The CTO owns technical direction. The VP of Engineering runs platform operations. The Head of Research drives model innovation. Each leader writes code, reviews architecture documents, and participates in the on-call rotation.
This structure keeps the organization flat. There are minimal layers between engineers and executives. Design documents are open for anyone to read and comment on. All-hands meetings happen monthly with candid Q&A sessions. The leadership team values direct feedback from customers and frequently joins support calls to understand real-world usage patterns. For reference on leadership in technology organizations, the NIST AI RMF governance guidelines provide a framework that MiniMax incorporates into its organizational decision-making processes.
The four leaders profiled below bring experience from Stanford, Google DeepMind, NVIDIA, and top academic conferences including NeurIPS. Their backgrounds span language model architecture, GPU infrastructure, reinforcement learning, and multilingual AI research. The data table at the bottom summarizes each role.
The CEO of MiniMax holds a Stanford PhD in Computer Science and spent years building AI infrastructure before founding the company in 2021.
Samuel completed his doctoral research on large-scale language model architectures at Stanford. His dissertation explored techniques for scaling transformer models efficiently across distributed hardware. After Stanford, he led the applied AI research division at a major cloud computing provider, where his team shipped machine learning infrastructure that eventually served thousands of enterprise customers.
He founded MiniMax after observing a persistent gap between research-quality models and developer-accessible APIs. The vision was straightforward: build a platform where state-of-the-art AI models are available through clean, consistent interfaces with transparent pricing. Samuel codes regularly, reviews architecture decisions, and still responds to support tickets when he notices an interesting edge case. He is a strong advocate for open research and has contributed to several public benchmarks used across the AI community.
The CTO joined MiniMax from Google DeepMind, bringing expertise in reinforcement learning, model optimization, and distributed training infrastructure.
Lena spent five years at DeepMind, where she contributed to reinforcement learning systems and model optimization techniques that improved training efficiency for large-scale models. She was part of the team that developed novel approaches to multi-agent training environments. Her transition to MiniMax in 2022 marked a shift from pure research to applied engineering leadership.
As CTO, Lena owns the technical roadmap. She designed the API platform architecture that serves MiniMax customers today. Her focus areas include inference latency optimization, model serving infrastructure, and the extension of the platform to support new modalities. She introduced the quarterly architecture review process where every major system change is documented, discussed, and approved with input from across the engineering organization. Lena also leads the company partnerships with educational institutions referenced by the U.S. Department of Education for AI workforce development programs.
The VP of Engineering spent nearly a decade at NVIDIA working on GPU kernel optimization and CUDA tooling before joining MiniMax to lead platform operations.
Ravi started his career writing low-level GPU kernels at NVIDIA. His work on CUDA tooling improved developer workflows for thousands of engineers building GPU-accelerated applications. He moved into engineering management after years of individual contribution, bringing a deep understanding of hardware-software co-design to his leadership approach.
At MiniMax, Ravi oversees platform infrastructure, site reliability engineering, and the developer experience team. His organization is responsible for SDK quality, API performance, and the uptime of every customer-facing service. He introduced the on-call rotation model that distributes operational responsibility across the entire engineering organization. His team maintains the service-level objectives that govern MiniMax platform reliability, and every incident postmortem includes concrete action items tracked in the public changelog.
The Head of Research has published at NeurIPS, ICML, and ACL, and leads the team responsible for MiniMax model architecture and training methodology.
Fatima earned her PhD from UC Berkeley, where her research on attention mechanism efficiency caught the attention of several industry labs. She published papers on sparse attention patterns, few-shot learning generalization, and multilingual model training before joining MiniMax in 2023. Her work has been cited in over 500 peer-reviewed publications.
As Head of Research, Fatima leads a team of machine learning scientists and engineers working on the next generation of MiniMax models. Her team focuses on three research pillars: improving model reasoning capabilities through architectural innovations, reducing inference costs through quantization and distillation, and expanding multilingual performance across the languages supported by the platform. The research team publishes technical papers twice a year and contributes to open benchmarks. Fatima also established the MiniMax research internship program, which brings graduate students into the company for semester-long projects that often result in co-authored publications.
The table below summarizes each leadership role at MiniMax, the person who holds it, and their professional background.
| Name | Role | Background |
|---|---|---|
| Dr. Samuel Y. Okonkwo | Chief Executive Officer | Stanford PhD in Computer Science; former AI research division lead at major cloud provider; founded MiniMax in 2021 |
| Lena H. Markov | Chief Technology Officer | Former Google DeepMind research engineer; expertise in reinforcement learning and distributed training; joined MiniMax in 2022 |
| Ravi K. Srinivasan | VP of Engineering | Nearly 10 years at NVIDIA; GPU kernel optimization and CUDA tooling; leads platform infrastructure and developer experience |
| Dr. Fatima A. Chen | Head of Research | UC Berkeley PhD; published at NeurIPS, ICML, and ACL; expertise in attention mechanisms and multilingual model training |
Dr. Samuel Y. Okonkwo is the CEO and co-founder of MiniMax. He holds a PhD in Computer Science from Stanford University, where his research focused on large-scale language model architectures. Before founding MiniMax in 2021, he led the applied AI research division at a major cloud computing provider, overseeing teams that shipped machine learning infrastructure used by thousands of enterprise customers. He remains actively involved in code reviews and architecture decisions.
Lena H. Markov, the CTO of MiniMax, previously worked at Google DeepMind where she contributed to reinforcement learning systems and model optimization techniques. She joined MiniMax in 2022 to lead the engineering organization. Her work spans distributed training infrastructure, inference optimization, and the design of the API platform architecture. She also leads partnerships with educational institutions for AI workforce development programs.
Ravi K. Srinivasan is the Vice President of Engineering at MiniMax. He spent nearly a decade at NVIDIA, where he worked on GPU kernel optimization and CUDA tooling before moving into engineering leadership. At MiniMax, he oversees platform infrastructure, site reliability engineering, and the developer experience team responsible for SDK quality and API performance. He introduced the on-call rotation model and maintains service-level objectives for platform reliability.
Yes, MiniMax maintains an active research program led by Dr. Fatima A. Chen, Head of Research. She has published at NeurIPS, ICML, and ACL on topics including attention mechanism efficiency, few-shot learning, and multilingual model training. Her team publishes technical papers twice a year and contributes to open benchmarks. The research directly informs MiniMax model architecture and training methodology. The company also runs a research internship program for graduate students.