Pontus is a privacy-first AI infrastructure platform that helps businesses build and deploy AI models more quickly and securely. It does this by providing a number of features that help businesses to manage and mitigate the risks associated with AI, including:

  • Automated data anonymization: Pontus can automatically anonymize data before it is fed into AI models, helping to protect the privacy of individuals.
  • Model risk management: Pontus provides a variety of tools for monitoring and managing the risks associated with AI models, such as bias, fairness, and security.
  • Zero-trust security: Pontus is built on a zero-trust security model, which means that it assumes that all users and services are untrusted by default. This helps to protect AI models from unauthorized access and modification.

 

Pontus also offers a number of other features that make it a valuable tool for businesses that are building and deploying AI models, such as:

  • Performance optimization: Pontus can help businesses to optimize the performance of their AI models, making them faster and more efficient.
  • Scalability: Pontus is scalable to meet the needs of businesses of all sizes, from startups to large enterprises.
  • Ease of use: Pontus is easy to use, even for businesses with no prior experience with AI.

Overall, Pontus is a powerful AI infrastructure platform that can help businesses to build and deploy AI models more quickly, securely, and efficiently. It is a valuable tool for businesses of all sizes that are looking to leverage the power of AI to improve their operations and achieve their business goals.

Chris: Hi there, please introduce yourself.

Sumant: Hi, I’m Sumant Shringari, the Co-founder and CEO of Pontus – the AI Privacy Layer.

Chris: Can you tell us more about your background and how did you come up with the idea to create Pontus?

Sumant: I am recovering academic, with degrees from the University of Pennsylvania, including a BA in Physics, Biophysics, and Biochemistry, an MS in Chemistry, and a Master’s in Computer Science. My academic journey was filled with deep dives into artificial intelligence and biophysics. But I eventually ventured out of academia and landed in the startup world.

At Gordian Software (YC W’19), I worked on the nitty-gritty, creating APIs that connected airlines and online ticketing agents. It was there that I crossed paths with my co-founder, Rushabh Mehta, and we clicked like two peas in a pod. Together, we created Seating Choice and took it from ground zero to raking in hundreds of thousands in ARR.

Our lightbulb moment for Pontus came after numerous conversations with industry leaders. They were all concerned about trust and privacy issues when dealing with these systems. The public and regulators were concerned about LLMs, making it clear that privacy and trust will be a major concern. That’s when the idea for Pontus hit us – a privacy layer for every interaction with an LLM.

Chris: What are the main features of Pontus?

Sumant: Pontus helps you use AI with privacy. We achieve this in three simple steps.

  1. Smart Anonymization. Pontus will take any prompt or context sent to LLM and redact information such as names, addresses, ssn numbers, and more. We do this through a NLP algorithm. Crucially, we redact intelligently, where we let the LLM know what was that info (ie John -> person_0), so the LLM can still personalize the.
  2. Auditing. Trust needs to be proven, so we create audit logs for every interaction done with an LLM. Regulators and users will ask for evidence in case of an audit, so we provide that out of the box.
  3. Zero Trust. Our system never stores sensitive information at rest or when interacting with third parties. In fact, you can even run Pontus on premise to fully own your data.

Chris: How does Pontus measure the effectiveness of its privacy layer?

Sumant: Pontus assesses the effectiveness of its privacy layer by utilizing default benchmark sets that evaluate how well Pontus anonymizes critical Personally Identifiable Information (PII) such as names, US addresses, Social Security Numbers (SSNs), and other sensitive data. This helps ensure the robustness of our privacy protection measures.

Chris: How does Pontus handle intellectual property rights? For example, if an organization uses Pontus to generate text that is copyrighted or trademarked, who owns those rights?

Sumant: The organization owns the rights. We make no claim to it.

Chris: Does Pontus offer any educational resources or training programs to help organizations use its product effectively?

Sumant: Yes, Pontus provides educational resources and training programs to support organizations in effectively utilizing our product. We offer implementation services and dedicated project management to ensure your success.

Chris: How do you ensure that Pontus is used ethically and safely?

Sumant: Pontus ensures ethical and safe usage by default through data anonymization. We’ve embedded privacy measures into our technology, making data anonymization the standard practice when using Pontus.

Chris: How can Pontus be integrated with existing enterprise systems?

Sumant: Integrating Pontus with existing enterprise systems is straightforward thanks to our on-premise self-hosted solution. This allows for seamless integration, as it operates on servers and provides an API structure that closely mirrors OpenAI’s APIs.

Chris: Who is your ideal customer for Pontus? How do you differentiate Pontus from your competitors?

Sumant: Our ideal customers for Pontus are mid-market companies and enterprises, particularly those where data privacy and trust are top priorities, as evidenced by conversations with industry experts.

Pontus sets itself apart from competitors by enabling organizations to move faster while prioritizing privacy. Unlike other privacy solutions that may hinder development, Pontus offers seamless integration and removes the concerns surrounding data privacy, allowing businesses to build and innovate without compromise.

Chris: Thanks for being with me, any last words? Where can our readers follow you?

Sumant: Thanks for the interview. You can follow me on Twitter and check out our company’s Linkedin page.