The workflow before#
Varde's investment team, managing about $13B, built its credit and cash flow models in R and Excel. Complex Monte Carlo simulations took 4 to 6 hours, large datasets crashed Excel, and overnight batch jobs delayed morning analysis. Each deal got only a few hundred scenarios, and results were hard to reproduce because models lived in individual analysts' files.
What we built#
We built a cloud analytics environment on Snowflake and AWS SageMaker. Market data (Bloomberg, Refinitiv), the portfolio management system, accounting and third-party research feed a single Snowflake warehouse.
We converted the core Excel and R models to Python and run them alongside statistical and machine learning models in SageMaker: Monte Carlo cash flow simulation, default prediction, credit spread analysis and portfolio risk. Simulations that took hours now finish in 5 to 10 minutes, and each analysis runs more than 10,000 scenarios.
We trained the investment team on Python, SageMaker and Snowflake, so analysts own and extend the models themselves.
Before and after#
| Measure | Before | After |
|---|---|---|
| Complex model runtime | 4 to 6 hours | 5 to 10 minutes |
| Scenarios per analysis | 100 to 500 | 10,000+ |
| Default prediction accuracy (AUC) | Baseline | 15% higher |
| Return forecast error (MAE) | Baseline | 20% lower |
What's next#
With models and data in one governed environment, the next step is Level 3 (Workflow Agents): an agent that assembles the first draft of the diligence pack from the models and data, with an analyst reviewing exceptions.
