Why we invested in Aporia, the observability platform for AI

Deploying machine learning algorithms in the wild can be risky business. Data scientists often manage to get the results they want when training algorithms in a controlled environment. But when those models go into production, external variables can impact performance, efficacy, security, and even outcomes.

Faulty outcomes and other deficiencies in machine learning models may not be apparent for weeks and sometimes even months. Relying on problematic artificial intelligence (AI) can result in adverse consequences. In order to ensure commercial readiness, organizations using AI must be sure that their machine learning models are accurate as they evolve over time.

Now, a customizable platform for monitoring machine learning applications enables users to ensure that their AI models are performing as expected. The innovative, full-stack platform developed by Aporia empowers data science teams to monitor, debug, explain, and improve their machine learning models in production. The platform is already being used by Fortune 500 companies in multiple industry sectors. In fact, over the past six months, Aporia has experienced 600% growth.

That growth helped Aporia raise $25 million in a Series A funding round led by Tiger Global, with participation from Samsung Next, and from TLV Partners and Vertex Ventures, both existing investors. We invested in Aporia because its team has proven that they have a go-to-market strategy that meets an important need in the marketplace.

Aporia’s founders have the experience needed to develop and commercialize this new observability tool for machine learning. Liran Paul Hason, CEO, is a former machine learning architect who worked with Adallom, a cloud security startup acquired by Microsoft. He also worked as an investor at Vertex Ventures, and previously served in the cyber intelligence unit of the Israeli Defense Forces (IDF). Alon Gubkin, CTO, worked as a developer at Dooble, Rakki, and Anzai – and was a research and development team lead in the IDF’s cyber intelligence unit.

The Aporia platform enables users to oversee their AI applications by providing full visibility into how machine learning models are performing in the real world. Data scientists can use the platform to create custom monitoring within minutes to detect a wide range of issues, such as biased predictions, unexpected changes in the format of the input data, or degradation in a model's performance over time. The platform also provides explainability and actionable insights that can help with troubleshooting any issue. Aporia designed its technology to seamlessly integrate with existing machine learning infrastructure.

Aporia’s product-led, go-to-market strategy includes a machine learning solution that can be integrated into a company’s development process, and used in staging and production. It monitors machine learning models in much the same way that application performance monitoring tools are used to help manage cloud apps. Because Aporia’s network agnostic platform is plug-and-play, it is easy to deploy with no runtime overhead.

We think Aporia’s technology is a proven cloud-based solution for enabling companies to monitor and explain their AI applications. In addition to meeting an immediate market need, Aporia provides a proactive way for companies to prepare for new regulations, such as the European Union’s impending new rules for monitoring the use of AI.

The time is right for a monitoring solution that empowers businesses to trust their AI. Aporia’s no code platform provides tools for visibility, monitoring and automation, explainability, and root cause analysis. Businesses that have embraced AI can now put safeguards in place to ensure that their machine learning models are safe, predictable, and profitable. 

Royi Benyossef is Investment Director at Samsung Next. Samsung Next's investment strategy is limited to its own views and does not reflect the vision or strategy of any other Samsung business unit, including, but not limited to, Samsung Electronics.

If you’re a founder, we’d like to meet you.