AI LEAD, JFROG
Somewhere between "AI will change everything" and "ok but who's going to actually build and deploy this."
18 years in enterprise tech. I lead AI on JFrog's CIO team. Day to day that looks like evaluating tools, building internal platforms, and working with engineering and security to get AI into actual workflows. Less "art of the possible," more "will this pass a security review."
I got here through a long stretch at Avaya, where I architected solutions for Dell, Apple, Boeing, and Comcast, and built tools for speech analytics and knowledge management. That background is probably why I care more about integration headaches than hype cycles.
I also mentor people on AI adoption and build things on the side, mostly productivity tools and automation.
Hebrew PII detection and redaction model.
No dedicated Hebrew PII detection model existed for local deployment. GolemPII fills that gap with a fine-tuned model trained on 115K Hebrew examples, enabling privacy compliance with Israeli data protection regulations without sending data to external APIs.
Cost visibility and session analytics for Claude Code.
Claude Code users had no way to track spending or analyze usage patterns across sessions. Claudoscope gives them a native macOS dashboard for real-time cost monitoring.
Align resumes with job descriptions for better ATS pass rates.
Job seekers struggle to tailor resumes for each application. Adapt-CV analyzes the gap between a resume and a job description, then suggests targeted improvements to increase ATS compatibility.
250K synthetic customer personas across six industries, ready for training and evaluating customer service AI systems.
Comprehensive Hebrew language dataset for training and evaluating PII detection and masking models. Contains 115K synthetic text samples representing various document types and communication formats commonly found in Israeli professional and administrative contexts.
What are you building?
I like connecting with people who ship things. Say hi.