Yariv Fishman is Chief Product Officer (CPO) at Deep Intuition, he is a seasoned product administration government with greater than 20 years of management expertise throughout notable international B2B manufacturers. Fishman has held a number of outstanding roles, together with management positions with Microsoft the place he led the Cloud App Safety product portfolio and initiated the MSSP and safety associate program, and Head of Product Administration, Cloud Safety & IoT Safety at CheckPoint. He holds a B.Sc in Info Techniques Engineering from Ben Gurion College and an MBA from the Technion, Israel Institute of Know-how.
Deep Intuition is a cybersecurity firm that applies deep studying to cybersecurity. The corporate implements AI to the duty of stopping and detecting malware.
Are you able to inform us about your journey within the cybersecurity trade and the way it has formed your method to product administration?
All through my 20 yr profession, I’ve labored at a number of international B2B organizations, together with Examine Level Software program Applied sciences and Microsoft, the place I led product administration and technique and constructed my cybersecurity expertise throughout public cloud, endpoint, community, and SaaS software safety.
Alongside the best way, I’ve discovered totally different greatest practices – from the right way to handle a crew to the right way to inform the right technique – which have formed how I lead at Deep Intuition. Working for quite a few cybersecurity corporations of assorted sizes has allowed me to get a holistic view of administration types and discover ways to greatest create processes that help fast-moving groups. I’ve additionally seen first-hand the right way to launch merchandise and plan for product-market match, which is important to enterprise success.
What drew you to hitch Deep Intuition, and the way has your position advanced because you began as Chief Product Officer?
As an trade veteran, I hardly ever get enthusiastic about new expertise. I first heard about Deep Intuition whereas working at Microsoft. As I discovered concerning the prospects of predictive prevention expertise, I shortly realized that Deep Intuition was the actual deal and doing one thing distinctive. I joined the corporate to assist productize its deep studying framework, creating market match and use circumstances for this first-of-its-kind zero-day information safety answer.
Since becoming a member of the crew three years in the past, my position has modified and advanced alongside our enterprise. Initially, I targeted on constructing our product administration crew and related processes. Now, we’re closely targeted on technique and the way we market our zero-day information safety capabilities in right now’s fast-moving and ever-more-treacherous market.
Deep Intuition makes use of a singular deep studying framework for its cybersecurity options. Are you able to talk about some great benefits of deep studying over conventional machine studying in risk prevention?
The time period “AI” is broadly used as a panacea to equip organizations within the battle towards zero-day threats. Nevertheless, whereas many cyber distributors declare to convey AI to the combat, machine studying (ML) – a much less subtle type of AI – stays a core a part of their merchandise. ML is unfit for the duty. ML options are skilled on restricted subsets of accessible information (sometimes 2-5%), supply solely 50-70% accuracy with unknown threats, and introduce false positives. In addition they require human intervention as a result of they’re skilled on smaller information units, rising the possibilities of human bias and error.
Not all AI is equal. Deep studying (DL), essentially the most superior type of AI, is the one expertise able to stopping and explaining recognized and unknown zero-day threats. The excellence between ML and DL-based options turns into evident when analyzing their potential to establish and forestall recognized and unknown threats. Not like ML, DL is constructed on neural networks, enabling it to self-learn and practice on uncooked information. This autonomy permits DL to establish, detect, and forestall complicated threats. With its understanding of the basic parts of malicious recordsdata, DL empowers groups to shortly set up and preserve a strong information safety posture, thwarting the subsequent risk earlier than it even materializes.
Deep Intuition lately launched DIANNA, the primary generative AI-powered cybersecurity assistant. Are you able to clarify the inspiration behind DIANNA and its key functionalities?
Deep Intuition is the one supplier available on the market that may predict and forestall zero-day assaults. Enterprise zero-day vulnerabilities are on the rise. We noticed a 64% improve in zero-day assaults in 2023 in comparison with 2022, and we launched Deep Intuition’s Synthetic Neural Community Assistant (DIANNA) to fight this rising pattern. DIANNA is the primary and solely generative AI-powered cybersecurity assistant to offer expert-level malware evaluation and explainability for zero-day assaults and unknown threats.
What units DIANNA other than different conventional AI instruments that leverage LLMs is its potential to offer insights into why unknown assaults are malicious. As we speak, if somebody needs to elucidate a zero-day assault, they need to run it by a sandbox, which may take days and, ultimately, will not present an elaborate or targeted rationalization. Whereas worthwhile, this method solely affords retrospective evaluation with restricted context. DIANNA would not simply analyze the code; it understands the intent, potential actions, and explains what the code is designed to do: why it’s malicious, and the way it would possibly influence programs. This course of permits SOC groups time to deal with alerts and threats that really matter.
How does DIANNA’s potential to offer expert-level malware evaluation differ from conventional AI instruments within the cybersecurity market?
DIANNA is like having a digital crew of malware analysts and incident response specialists at your fingertips to offer deep evaluation into recognized and unknown assaults, explaining the methods of attackers and the behaviors of malicious recordsdata.
Different AI instruments can solely establish recognized threats and present assault vectors. DIANNA goes past conventional AI instruments, providing organizations an unprecedented degree of experience and perception into unknown scripts, paperwork, and uncooked binaries to organize for zero-day assaults. Moreover, DIANNA offers enhanced visibility into the decision-making technique of Deep Intuition’s prevention fashions, permitting organizations to fine-tune their safety posture for max effectiveness.
What are the first challenges DIANNA addresses within the present cybersecurity panorama, notably relating to unknown threats?
The issue with zero-day assaults right now is the lack of knowledge about why an incident was stopped and deemed malicious. Menace analysts should spend important time figuring out if it was a malicious assault or a false constructive. Not like different cybersecurity options, Deep Intuition was routinely blocking zero-day assaults with our distinctive DL answer. Nevertheless, prospects have been asking for detailed explanations to higher perceive the character of those assaults. We developed DIANNA to reinforce Deep Intuition’s deep studying capabilities, cut back the pressure on overworked SecOps groups, and supply real-time explainability into unknown, subtle threats. Our potential to focus the GenAI fashions on particular artifacts permits us to offer a complete, but targeted, response to handle the market hole.
DIANNA is a big development for the trade and a tangible instance of AI’s potential to resolve real-world issues. It leverages solely static evaluation to establish the conduct and intent of assorted file codecs, together with binaries, scripts, paperwork, shortcut recordsdata, and different risk supply file varieties. DIANNA is greater than only a technological development; it is a strategic shift in the direction of a extra intuitive, environment friendly, and efficient cybersecurity setting.
Are you able to elaborate on how DIANNA interprets binary code and scripts into pure language stories and the advantages this brings to safety groups?
That course of is a part of our secret sauce. At a excessive degree, we are able to detect malware that the deep studying framework tags inside an assault after which feed it as metadata into the LLM mannequin. By extracting metadata with out exposing delicate data, DIANNA offers the zero-day explainability and targeted solutions that prospects are in search of.
With the rise of AI-generated assaults, how do you see AI evolving to counteract these threats extra successfully?
As AI-based threats rise, staying forward of more and more subtle attackers requires transferring past conventional AI instruments and innovating with higher AI, particularly deep studying. Deep Intuition is the primary and solely cybersecurity firm to make use of deep studying in its information safety expertise to stop threats earlier than they trigger a breach and predict future threats. The Deep Intuition zero-day information safety answer can predict and forestall recognized, unknown, and zero-day threats in <20 milliseconds, 750x quicker than the quickest ransomware can encrypt – making it an important addition to each safety stack, offering full, multi-layered safety towards threats throughout hybrid environments.
Thanks for the nice interview, readers who want to study extra ought to go to Deep Intuition.