Skip to content

Aethel Economic Scenario Generator (ESG)

Aethel is an actuarial-grade Economic Scenario Generator (ESG) engine written in Python. It implements high-performance continuous-time stochastic processes to generate correlated multi-path simulations of core macroeconomic indicators and asset classes.

Primary Applications

Aethel is optimized for modeling long-term economic trajectories, helping risk managers and financial planners conduct:

  • Asset-Liability Management (ALM) and solvency evaluations.
  • Retirement Decumulation Profiling (Sequence of Returns Risk modeling).
  • Dynamic Portfolio Projections under varying inflationary and interest rate regimes.
  • Yield Curve Term-Structure Forecasting (Nominal and Real).

Aethel Cloud API (SaaS Platform)

While the mathematical modeling engine can be executed locally, we offer a fully managed Multi-Tenant SaaS API designed to transition stochastic simulations from local scripts to scalable, cloud-hosted architecture.

The API layer wraps the Aethel core engine in a production-grade Web Service built with FastAPI, hosted on Render, and backed by a Neon (PostgreSQL) database for persistent tenant, billing, and secure api-key management.

Cloud-Only Features & Enterprise Capabilities

  • Premium Pre-Calibrated Presets: Run simulations instantly using built-in, historically calibrated macroeconomic baselines for major global regions, including USA, Europe, Japan, Brazil, and World.
  • Expert Actuarial Boundaries: Regional presets include structural overrides (such as deflation floors and nominal rate boundaries) designed to preserve simulation stability in extreme interest rate regimes.
  • Enterprise Calibration Lifecycle: Upload raw historical CSV datasets containing interest rates, inflation, and equity returns directly via the /api/v1/calibrate endpoint to fit parameters on cloud hardware and save them as permanent, custom presets.

  • Resource & Quota Management: Compute usage is tracked automatically using a robust Compute Unit billing metric, derived as:

    \[ \text{Compute Units} = \text{Scenarios} \times \text{Projection Years} \times 12 \]

  • Tenant Access Control: Subscription tiers (Free, Professional, Enterprise) enforce API rate limits and monthly cumulative compute budgets. Authentication is secured using crytographically hashed SHA-256 API keys.

Launch Live Interactive Dashboard | Request API Access via Tally

High-Performance Architectural Features

To manage the high memory demands of multi-scenario simulations, the engine implements several performance optimizations:

  • Hardware-Aware Concurrency: The simulation detects available host RAM and CPU capacity to automatically partition simulations into memory-safe blocks, preventing page-fault issues.
  • Memory-Safe Lazy Evaluation: Generating term structures (nominal and real yields) for multiple maturities across thousands of scenarios is memory-intensive. The engine uses a LazyScenarioList to keep primary arrays contiguous in memory, deriving yield curves dynamically only when accessed.
  • Fastmath Compilations: If numba is detected, the engine runs JIT-compiled loops with parallel execution capabilities.

Installation

# Clone the repository
git clone https://github.com/Gustavo1500/aethel-esg.git

cd aethel-esg

# Install the package locally with visualization and optimization dependencies
pip install -e ."[numba,plots]"