Amir Shachar
Fraud & Cyber Scientist
Founder and Chief Scientist at Riskernel. Previously, served as Chief AI Scientist at Skyhawk Security, directing AI-driven threat detection research, and before that, headed Fraud Data Science at NICE Actimize, deploying fraud models across the world’s largest banks. Here I write about the math behind these systems, and about AI that proposes before it acts.
- Before
- Former Head of Fraud Data Science, NICE Actimize
- Patents
- Named inventor on 15 patents, 9 granted in the US
Contents
Fraud data science
I write about the math of catching fraud, with the data and the failure case shown.
- Fraud math: base rates, class imbalance, drift
- Models that propose, validators that decide
- Learning across institutions
- Attack anatomy, from public sources
Latest posts
- Alert budgets, not AUC: judging a fraud model in the units that matter Fraud is rare and review capacity is fixed, so judge a fraud model by the frauds in the alerts a team can open and by the money each decision moves.
- Fraud rings and mule networks: what a graph adds, and what it cannot decide Ring payments look normal one at a time, so the signal is in shared devices, beneficiaries and timing, and a graph score is a lead to confirm.
- Propose, then verify: where a language model belongs in a fraud stack A language model can propose fraud rules and explanations, but a deterministic check, separate from the model, should decide what runs in a payment flow.
Fig. 1 · Where a fraud team works on the ROC curve
Paper, 2024, with a fraud preface
Algogens
A framework in which generative models propose and deterministic validators decide what runs. I first set it out in a 2024 paper, and Riskernel's engine is built on it.
Propose A model proposes, for example a new feature, a new rule or a mapping for a new data feed.
Validate A deterministic validator checks every proposal before any action runs.
Repair or reject A failed proposal gets a bounded repair, or it is rejected.
Act Only a proposal that passed runs.
I apply this at Riskernel, the fraud engine I founded. riskernel.com ↗
Mathematics
Semi-discrete Calculus and the pages around it.
About
Hello and welcome! I’m passionate about the intersection of mathematics, science, and technology. My journey as a software developer led me to the fascinating world of AI, a field I’ve come to love.
I lead a data science consulting organization that also extends pro-bono expertise to charitable causes, using data to make a real difference.
I enjoy storytelling through science, inspired by educators like Sabine Hossenfelder, Vi Hart, Grant Sanderson, Michael Stevens, Sal Khan, and Lex Fridman. Their approach to making complex concepts accessible and engaging deeply influences my work.
My research focuses on Semi-discrete Calculus, Algogens, and AI, offering endless opportunities for discovery and innovation.
Curious to learn more about my journey? Visit my LinkedIn profile for a detailed glimpse into my professional world.
If you’re interested in collaborating or chatting about AI and science, feel free to reach out via email. Let’s connect and explore the possibilities together!