Where machine
learning meets
financial practice
Vephoxar Lre started in 2022 with a straightforward premise: most people working in finance encounter ML tools without understanding how they actually function. That gap creates risk. We built a workshop platform specifically to close it — through hands-on exercises, not slide decks.
How the platform was shaped by real friction
Our instructors spent years inside financial institutions watching analysts misapply regression models, misread feature importance scores, and trust black-box outputs without validation. The workshops here are built around those specific failure points — not hypothetical ones.
Each assignment mirrors a real scenario: credit risk scoring with imbalanced datasets, anomaly detection in transaction logs, forecasting with noisy time-series data. Participants work through these problems with actual tools — Python, scikit-learn, pandas — not simplified toy environments.
Who teaches here — and what that means for you
Instructors at Vephoxar Lre are practitioners first. Tamar Eshkoli leads quantitative risk modules after eight years building credit models at a regional bank. Yoav Drori covers algorithmic pattern recognition, drawing from his background in institutional portfolio analytics.
Neither of them teaches from textbooks alone. Sessions are structured around problems they personally had to solve — which means the edge cases, the gotchas, and the debugging process all make it into the curriculum.
Recent platform activity
The platform runs on a rolling workshop schedule. Cohorts stay small — typically 12 to 18 participants — so instructors can give meaningful feedback on individual submissions rather than generic comments.
-
New cohort opened — Credit Scoring with Gradient Boosting module. 4 spots remaining.
-
Assignment review completed — Transaction Anomaly Detection batch. Feedback sent to 16 participants.
-
Live session recorded — Feature Selection in Noisy Financial Data. Available to enrolled participants.
-
Updated exercise set — Time-Series Forecasting module now includes LSTM comparison exercises.