Shay Levi is the co-founder and CEO of Unframe, a company with enterprise AI with scalable, secure solutions. Previously, he co-founded non-Amam Security and led the company to a $500 million acquisition through Akamai in just four years. He is a proven innovator in the field of cybersecurity and technology, specializing in building transformative solutions.
Unframe is an all-in-one enterprise AI platform based in Cupertino, California, enabling businesses to bring any unique AI use cases to their lifespan in hours rather than months. Through its Blueprint approach, Unframe works with large global enterprises to implement solutions across observation, data abstraction, smart proxy and modernize. Unframe uses results-based pricing, allowing customers to experience and develop whatever solution they want without risk. UNFRAME is LLM agnostic and does not require fine-tuning or training – essentially changing the possibilities of large enterprises seeking scalable Turnkey AI solutions.
On April 3, 2025, Unframe appeared $50 million from stealth to change enterprise AI deployment.
After successfully exiting Akamai, what prompted you to launch Unframe and what gaps did you identify in the enterprise AI space to make it the right time and opportunity?
Actually, I actually left Noname before the acquisition discussion began. What I see is a huge wave, with CIOs under pressure to adopt AI quickly, but the tools they have available are not ready for businesses. They either piece together point solutions without governance or wait for the internal team to build from scratch. Neither path is scaled and both introduce risks.
That’s a signal. I realized that businesses need not only access to AI – they need a platform, while providing them with control, speed and flexibility. This is what causes Unframe.
NONAME SECurity is a pioneer in API network security. How did your experience building a security-centric company shape the approach you take with Unframe?
Safety is in our DNA. At Noname, we learn that innovation without governance quickly leads to risks. The course continues directly to AI. From day one of Unframe, we’ve baked in the right guardrails – secure data processing, model transparency, role-based access – so businesses can innovate without introducing new vulnerabilities.
We are also very aware of the massive breakdown of things. So while Unframe gives teams a quick action, we have already considered the platform for enterprise complexity – whether it is managing data flows, performing compliance, or integrating with Legacy Systems.
Are there any common pain points among enterprises in the API security space that help inform you about your vision for AI adoption?
really. At NONAME, we see the challenge of businesses gaining visibility and control in their environments. Shadow APIs, inconsistent tools and siloed teams create real operational risks and slows down everything.
Using AI, we see the same pattern repeating. Every team wants to move quickly, but without the right structure, you get splits, repetitions and blind spots. This experience shapes our thinking with Unframe. We want to provide businesses with a way to adopt AI in a unified, secure and practically work on teams and systems, rather than just isolated pockets.
UNFRAGE gained the appeal of major businesses in one year and received ARR in millions of dollars – how did you reach this level of adoption so quickly?
We did not take traditional slow experimental pathways or limited pilots. From day one, we have been working in the market, working with global businesses in high-impact real-world projects. These are not siloed use cases – they are strategic initiatives related to the core part of the business. That’s why we win America’s trust and help Unframe become a strategic partner across multiple fields, not just suppliers. It adopts when you provide real results quickly.
You’ve talked about reducing AI deployment from months to hours. Can you take us through how UNFRAME does this?
We have built hundreds of deep technology building blocks in the Unframe platform. When deploying new solutions, it is not from scratch – it is tailored by mapping these components to blueprints for user-specific needs. That’s how we’re reducing deployment from months to hours.
Tell us more about the Blueprint approach – what makes it unique and why it is so powerful for enterprise AI, for example?
The Blueprint approach is how we deliver tailor-made AI solutions at scale – without having to start from scratch. Each blueprint draws logic, components, workflows, and guardrails for specific use cases and configures our platform’s library of technical building blocks. This is how we combine speed and accuracy.
Unframe positions itself as LLM-AGNOSTIC, and does not require fine-tuning of the model. Why is it important for you to avoid the need to train custom models?
Because most businesses do not need custom models, custom results are required. When you start fine-tuning, you lock yourself in a specific vendor, increase costs, and create maintenance overheads that most organizations can’t handle.
We designed UNFRAME to work with existing modern models to still deliver high-quality results tailored – without complexity. By maintaining LLM-Agnostic, we provide businesses with flexibility, faster value time and the ability to grow as the model landscape changes. The goal is not to train the model, but to solve the problem. And, you can do this well without touching fine-tuning.
What role does natural language interaction play in the Unframe platform and how far can it go in replacing traditional software UI?
Natural Language is a powerful entry point – it gives enterprise users immediate access to AI without training or technology upgrades. This is especially important when you work with companies around the world and distribute your workforce in different countries, roles and languages. Chat-style interface helps level the playing field.
However, each Unframe use case is different and the interface needs to match tasks. Sometimes this means natural language chat. Other times, this is a dynamic table, interactive dashboard, or content generation interface – whatever is best suited for the workflow and the results we want to solve.
We do not consider natural language to be a replacement for traditional UI, but rather a layer that eliminates important friction. The purpose is to make the software feel intuitive, flexible and tailored – not only for users, but also to solve the problems they are trying to solve.
What courses will you apply for at UNFRAGE, from scaling non-AME Security to $1B+ valuations and $450 million acquisitions?
Focus on solving real urgent issues – and use enterprise-level execution from day one. At Noname, we understand that scale comes from trust, and trust comes from rapid delivery without cutting corners. At Unframe, we are applying the same mindset: move quickly, build securely and stay relentlessly customer-centric.
As a repeat founder, what is your approach to building a leadership team and company culture in a hypergrowth environment?
In super growth, you don’t have a luxury to figure things out slowly – so you need people around you who are not only good at their jobs, but also need to thrive in ambiguity and urgency. For me, building a leadership team starts with trust, clarity, and sharing values. Everyone has to be consistent where we are going and equally committed to having their own journey.
The culture is the same. It’s not a table tennis table – when things get tough, that’s how you decide. At Unframe, we have always been interested in building a culture of ownership, pace and honesty. We move quickly, listen hard, and we push each other to be better every day. This culture can not only survive hypergrowth—it can drive it.
Thanks for your excellent interview, and readers who hope to learn more should visit Unframe.
