Jungle-labs
We believe cybersecurity should move beyond fragmented visibility and disconnected alerts. Jungle-labs was founded to build systems that operate with more coherence: adaptive where they need to be, controlled where they must be, and useful when timing matters.
Adaptive at the edge. Coherent at the core. Unsentimental about what works.
Cyber threats are now lethal.
AI-powered cyber attacks force us to grasp a brand new reality — one where threats move at machine speed, adapt mid-operation, and target the seams humans can no longer watch alone.
Defense-in-depth is not enough anymore. It is no longer only about protecting the perimeter — it is about reacting offensively, with discipline, to silence the threat before it compounds.
Jungle-labs is built for that reality. Offense and defense on the same side, in the same system, under a single operator's command.
When a compromised model moves.
AI is moving into machines that sense, move, and act. When a model is compromised on a drone, a robot, or an industrial controller, the consequence isn't a leak — it's motion.
Jungle-labs researches security for the embodied domain: adversarial attacks on perception and sensors, model and policy integrity on physical platforms, and the trust boundary where reasoning becomes an actuator command.
Where attackers will hunt next, we're already looking.
Systems, not products.
Modern environments are not suffering from a lack of security tooling. They are suffering from fragmentation — too many surfaces, too much signal, and too little coherence between them.
Jungle-labs is shaped by a different premise. Security infrastructure should behave like operational infrastructure: observing continuously, relating signal intelligently, and supporting the humans who carry the weight of decisions under pressure.
Our architecture draws from how living systems function — adaptive at the edge, coherent at the core, and useful in the absence of perfect information. That philosophy sits under every component we build.
How we build
Restraint over hype
We prefer systems that earn trust through operational behavior rather than positioning.
Depth over surface
Real security work happens beneath dashboards. We build for the layer that matters.
Respect for the operator
Security teams are the most underrated infrastructure in the stack. We design around them.
Built by operators and researchers
Jungle-labs is an Israeli company, shaped by a small founding team with direct experience across the security, research, and infrastructure worlds.
Tony Voloshko
Tony is the Founder and CEO of Jungle Labs, developing biomimicry-inspired, AI-native security tools. A veteran cybersecurity consultant, he builds autonomous systems designed to think, adapt, and hunt.
LinkedIn
Dr. Lev Topor
Dr. Lev Topor is a co-founder of Jungle Labs, an author, and a cybersecurity expert who leverages his deep research into intelligence operations and cyber policy to build adaptive, AI-driven offensive security solutions.
LinkedIn
K.
K. is a co-founder and the CTO of Jungle Labs, an AI-optimization engineer who architects the model layer behind Canopy, Macaw, and Sloth — squeezing reasoning-grade performance out of models small enough to run on an endpoint. He has asked to remain anonymous for now.
Profile privateSerious buyers in serious environments
Jungle-labs is designed for organizations where security is not a feature — it is infrastructure. We work with teams who live inside the consequences.
CISOs
Leaders accountable for cyber outcomes — not just controls.
Security leaders
Heads of engineering, detection, and operations carrying the weight of decisions.
Security engineering
Practitioners who need tools that respect how real systems behave.
Enterprise buyers
Teams evaluating platforms that can operate in regulated and sensitive contexts.
Partner with Jungle-labs
We work closely with a small number of design partners who share our commitment to building serious cyber infrastructure.