Ameesh Divatia is the co-founder & CEO of Baffle, an organization targeted on integrating information safety into each facet of the information pipeline to simplify cloud information safety and reduce the influence of knowledge breaches.
Its platform presents a no-code, easy-to-deploy answer that secures delicate information with out affecting efficiency or requiring adjustments to purposes. Baffle’s know-how is appropriate with main cloud suppliers resembling AWS, Azure, IBM, and GCP. Serving a variety of purchasers, from Fortune 25 corporations to small and medium companies, Baffle protects over 100 billion data worldwide, working with system integrators for environment friendly deployment.
What motivated you to co-found Baffle, Inc., and the way did your earlier entrepreneurial experiences form your method within the early phases of the corporate?
After my final firm’s exit, I took a much-needed break to recharge and take into consideration what I actually wished to do subsequent. I’ve all the time liked constructing corporations, so I began having conversations with an early-stage VC good friend of mine, and he launched me to Priyadarshan “PD” Kolte, who would develop into my co-founder. He challenged us with an intriguing query, disguised as a problem: “How do you get value from data while still protecting it?” That problem hooked me—fixing robust issues is what I stay for. There was a obvious hole in information safety, particularly round simplifying encryption and defending information ‘in use’. 9 years later, right here we’re, answering that query with Baffle.
With the rise of generative AI, how can corporations be certain that their information stays safe whereas nonetheless leveraging the advantages of AI applied sciences?
It is a query each firm dabbling in AI ought to be asking. Safety and innovation typically really feel like two opposing forces, however they don’t need to be. The bottom line is a breakthrough innovation known as Privateness Enhanced Computation (PEC) that begins with encryption—holding information protected at relaxation, in transit, and whereas in use. By encrypting delicate information earlier than it will get to the AI fashions first after which utilizing PEC to course of it, you possibly can nonetheless get the insights you want with out compromising safety. It’s about staying forward of the sport, updating safety protocols, and leveraging instruments like Baffle to mitigate dangers. You don’t need to sacrifice innovation for safety.
Are you able to clarify the precise position of encryption in defending AI-generated information and fashions? How does it differ from conventional information safety strategies?
Encryption for AI information is like wrapping your Most worthy asset in bubble wrap—regardless of how a lot it’s tossed round, it stays protected. Consider it as locking the information whilst you’re utilizing it. Conventional strategies give attention to securing information when it’s not in use (at relaxation) or when it is transferring (in transit). However with AI, we’re including a brand new layer of complexity as a result of the information wants to remain encrypted even when it’s being crunched by fashions. Baffle focuses on this “data-in-use” safety, making certain efficiency isn’t impacted however safety isn’t sacrificed.
Baffle just lately launched a knowledge safety answer particularly for GenAI tasks. Are you able to share extra particulars about how this answer works and what makes it distinctive out there?
Our GenAI answer is all about making encryption easy and environment friendly, even if you’re working with AI. It plugs into an current AI pipeline by defending information as it’s being ingested. That is adopted by a functionality referred to as real-queryable encryption that processes the information with out exposing it. Most significantly, you don’t want to vary something in your AI pipeline—no rewriting code, no trouble. Simply plug it in and go. We’ve targeted on ease of use and ensuring safety doesn’t get in the way in which of innovation, which is why prospects are discovering this answer so engaging.
Your platform emphasizes “no code” adjustments for implementing information safety. How does this method profit corporations, particularly these with massive, advanced information pipelines?
Nobody needs to interrupt one thing that’s already working. With our “no code” method, corporations don’t want to tear aside their current purposes or information movers so as to add encryption. It is a big profit for big organizations with advanced information pipelines as a result of it means they will bolster safety with out risking disruptions. It’s sooner, simpler, and removes numerous the complications that sometimes include integrating new tech.
How does Baffle’s Actual Queryable Encryption differ from different encryption strategies, and what benefits does it provide for corporations dealing with large-scale information analytics?
Actual Queryable Encryption is our secret sauce. Not like conventional encryption, which requires you to decrypt information within the information retailer earlier than analyzing it (and thereby exposing it), we allow you to run queries on the encrypted information itself. It’s like having your cake and consuming it too—you get the insights with out risking safety. It is a game-changer, particularly for corporations coping with big quantities of delicate information, like in finance or healthcare, the place compliance is non-negotiable.
Information-in-use safety is a essential characteristic of Baffle’s platform. Are you able to clarify how this works and why it’s important for corporations, significantly within the context of GDPR and different information privateness rules?
When information is in use—being processed by methods—it’s normally at its most susceptible. That’s why defending it in real-time is essential, particularly with rules like GDPR, which require a posture referred to as ‘data protection by design’. Our platform ensures that even when information is being processed, it is nonetheless encrypted. This method eliminates that dangerous window of publicity the place information breaches typically occur, serving to corporations keep compliant and protected.
As AI fashions develop into extra advanced, what are the principle challenges in securing these fashions towards adversarial assaults, and the way does Baffle deal with these challenges?
AI fashions are getting smarter, however so are attackers. Adversarial assaults—the place unhealthy actors attempt to manipulate the information that impacts an AI mannequin’s output—are a rising concern. We deal with this by specializing in the information facet. By encrypting the information that the AI fashions depend on, we make it a lot tougher for anybody to mess with the mannequin’s integrity. It’s like giving the AI mannequin a locked vault of knowledge—nobody’s getting in with out the important thing.
Are you able to focus on the significance of role-based entry management (RBAC) in trendy information safety methods, particularly for organizations utilizing multi-tenant cloud environments?
In multi-tenant cloud environments, RBAC is a must have. Think about you’ve obtained a bunch of individuals all sharing the identical cloud infrastructure. With out RBAC, it’s like giving everybody entry to the entire constructing as a substitute of simply their workplace. Our platform integrates RBAC so solely approved individuals based mostly on their particular person position or credential, can entry delicate information, holding issues locked down tight and decreasing the danger of breaches.
Baffle has seen important progress lately, along with your revenues doubling previously yr. What do you attribute this progress to, and the way do you propose to proceed this trajectory?
We’re using a wave of demand as a result of we’ve constructed the precise answer for the precise drawback. Our progress comes down to at least one factor: we’re fixing an issue that each firm faces—information safety. With cyber threats on the rise and rules getting harder, corporations are turning to us for options that work with out slowing them down. Our give attention to real-queryable encryption and ease of use is a giant purpose for that progress. Shifting ahead, we’re planning to maintain pushing the envelope on innovation, increasing our merchandise, and constructing sturdy partnerships that take us into new markets.
Thanks for the good interview, readers who want to be taught extra ought to go to Baffle.
