Data Annotator

Description DLP is broken. Too many alerts, zero context, security teams drowning in noise. Jazz flips the model - full visibility from the endpoint, AI investigators that understand your business, and real answers instead of alerts. Built so one person can actually run it. Finally, security that works. At the core of that AI is training data. And training data is only as good as the people reviewing it. We're looking for a sharp, detail-oriented Data Annotator to be the quality checkpoint between raw security events and our ML pipeline - making sure the model learns from the right examples, every time. WHAT YOU'LL DO Review DLP events and AI-generated decisions, labeling them as accurate or inaccurate according to established guidelines, with precision and consistency at volume. Identify edge cases and ambiguous events and escalate them to the ML and Security team with a clear, reasoned explanation of the call. Help refine annotation guidelines over time as patterns emerge and the model evolves. Investigate the root causes of labeling decisions and develop an understanding of how LLM drift and feature importance affect model behavior. Maintain a high and consistent accuracy rate, whether it's your tenth or your thousandth review of the day. WHAT WE'RE LOOKING FOR Studying or recently graduated in CS, information systems, medicine, or a related field - or equivalent hands-on experience in a data, QA, or analytical role. Background in a security company, SaaS startup, or a military intelligence/analytical unit. Basic grasp of how ML/LLM models make decisions - you don't need to build them, but you should understand the logic behind them. Comfortable with dashboards and data tools, you follow guidelines precisely and flag when something doesn't fit. Reliable and detail-obsessed! You catch what others miss, and your accuracy holds up across hundreds of reviews, not just the first ten. Clear communicator. You can identify broader patterns and translate complex labeling decisions into straightforward language. Good judgment - you know when to decide and when to escalate. Prior experience with annotation tools like Label Studio or Scale AI is a plus. WHY JOIN JAZZ Direct, visible impact - the work you do today improves the core product. You'll see the results land in production and reach real customers. Exposure to cutting-edge AI security technology and a team that genuinely cares about getting it right. A foot in the door at a fast-growing cybersecurity startup, with a team that values people who grow into their roles. Our office is located at Sonol Tower, Tel Aviv, next to Carlebach rail station.