Robust Intelligence is a technology company that focuses on securing artificial intelligence systems. Founded in 2019 by Yaron Singer and Kojin Oshiba, the company provides a platform designed to validate and protect machine learning models from adversarial attacks, data drift, and other security vulnerabilities. Its headquarters are located in San Francisco, California.
The company emerged from research at Harvard University, where Singer is a professor of computer science and applied mathematics. Robust Intelligence aims to address the growing need for AI security in enterprise environments, particularly as organizations increasingly deploy machine learning models in critical applications such as fraud detection, autonomous vehicles, and healthcare diagnostics.
History and Founding
Robust Intelligence was co-founded in 2019 by Yaron Singer and Kojin Oshiba. Singer, who serves as CEO, brought academic expertise in algorithms and machine learning, while Oshiba, the CTO, contributed engineering and product development experience. The company initially operated in stealth mode, developing its core technology before publicly launching in 2021.
In 2021, Robust Intelligence announced a $30 million Series A funding round led by Sequoia Capital, with participation from other investors including Harrison Metal and Y Combinator. The funding was intended to accelerate product development and expand the company's go-to-market efforts. By 2023, the company had raised additional funding, bringing its total to over $60 million.
Platform and Technology
The core offering of Robust Intelligence is its AI Validation and Protection Platform. The platform continuously tests machine learning models for vulnerabilities, including adversarial examples, data poisoning, and out-of-distribution inputs. It uses a combination of automated red-teaming, stress testing, and monitoring to identify potential failure modes before they can be exploited.
The platform integrates with existing ML workflows and supports popular frameworks such as TensorFlow, PyTorch, and scikit-learn. It provides a dashboard for security teams to visualize risks, receive alerts, and generate compliance reports. The system also offers mitigation strategies, such as adversarial training and input sanitization, to harden models against identified threats.
Industry Context
The rise of artificial intelligence and machine learning has created new security challenges. Traditional cybersecurity focuses on protecting data and infrastructure, but AI models introduce unique vulnerabilities. Adversarial attacks can manipulate inputs to cause misclassification, while data drift can degrade model performance over time. Robust Intelligence operates in this niche, competing with other AI security startups and broader cybersecurity vendors.
Key related concepts include adversarial machine learning, data drift, and model validation. The company's approach is often compared to penetration testing for AI, as it proactively seeks out weaknesses rather than waiting for incidents to occur.
Leadership and Team
Yaron Singer, the CEO, is a prominent researcher in algorithms and machine learning. He has published extensively and has been recognized with awards such as the NSF CAREER Award and the Sloan Research Fellowship. Kojin Oshiba, the CTO, previously worked at Google and has experience in large-scale systems and product engineering.
The company has attracted talent from leading tech firms and research institutions, including former employees of Google, Microsoft, and various universities. The team includes experts in security, ML research, and software engineering.
Funding and Growth
Robust Intelligence has raised significant venture capital funding. The Series A round in 2021 was followed by a Series B round in 2023, which brought total funding to $60 million. Investors include Sequoia Capital, Harrison Metal, Y Combinator, and others. The company has used the funding to expand its sales team, enhance its product, and build partnerships with cloud providers such as Amazon Web Services and Google Cloud.
As of 2023, Robust Intelligence serves a range of enterprise clients across finance, healthcare, and technology sectors. The company reports that its platform has been used to test and protect thousands of models, though specific customer names are often kept confidential.
Future Directions
Looking ahead, Robust Intelligence aims to broaden its platform to cover the entire AI lifecycle, from development to deployment. The company is also exploring ways to address emerging challenges in large language models and generative AI, which present new types of risks such as prompt injection and biased outputs. As AI adoption continues to grow, the demand for robust security and validation is expected to increase, positioning the company for further expansion.
See Also
- adversarial machine learning
- data drift
- model validation
- AI security
- Yaron Singer