Yuval Herman
Senior Data Scientist, Tel Aviv, Israel
yuvalherm@gmail.com linkedin.com/in/hermanyuval yuvalherman.co.il
Profile
Data Scientist with 6 years of experience developing and deploying machine learning solutions, specializing in credit risk and fraud. Experienced in end-to-end modeling, statistical evaluation and applied AI, with a prior career in accounting and finance (CPA).
Experience
Data Scientist, Fido Credit Solutions, Tel Aviv 2020–Present
- Owned credit-risk modeling across three international markets, from feature development and model evaluation through deployment and monitoring, improving ROC-AUC by 0.07 (absolute) over baseline models.
- Developed a causal decisioning layer for Collections using uplift modeling, driving treatment decisions by each customer's estimated response.
- Built time-series regression forecasts for Finance using Prophet, modeling trend and seasonality.
- Developed and introduced multiple fraud-detection rules for the lending population, reducing fraud by 2.5%.
- Engineered behavioral and graph-based features into risk-model pipelines.
- Designed GenAI-automated workflows for model training, feature-accuracy validation and document parsing, built from reusable agent skills, agentic loops and human-in-the-loop review checkpoints.
Software Developer & Blockchain Administrator, Dov & Co, CPA 2018–2020
- Developed financial software and blockchain-related applications; advised on digital-asset regulation, compliance and tax.
Financial Auditor, UHY Israel 2015–2018
- Audited public and private companies and supported financial reporting; also served as controller at a listed financial-services firm.
Technical skills
- Programming
- Python, SQL, Gremlin
- Machine learning
- LightGBM, XGBoost, CatBoost, scikit-learn, PyTorch Geometric (GNNs), Prophet, Optuna, SHAP
- Statistics & experimentation
- A/B testing, Time-series forecasting, Causal inference, Uplift modeling, Imbalanced data, Explainability, Drift monitoring
- Applied AI
- LLM agents, RAG, Evaluations, MCP, AWS Bedrock, OpenAI
- MLOps & deployment
- MLflow, SageMaker, Docker, Git, Lambda, Step Functions
- Data & cloud
- AWS, Snowflake, Redshift, Neptune Graph
Selected projects
The Global Min: A data-science interview-prep app for iOS with 900+ adaptive scenario questions across nine topics. It finds your level per topic and moves you up. The question bank was built with LLM-assisted generation and automated quality checks. theglobalmin.app
Revaluation: Finance research applying convex optimization, yield-curve modeling and machine learning to bond-risk analysis. Source on GitHub
Education & certification
Data Science Fellowship, Israel Tech Challenge (ITC), 2020
Selective program in machine learning, deep learning, computer vision and NLP.
B.A., Business Management (Accounting & Statistics), The Academic College of Management, 2012
Certified Public Accountant (CPA), Israel
Languages
Hebrew (native), English (native)
Interests
Chess, competitive player in a local community league