← Mandates
Eldorado
Eldorado Quantitative Research Platform
January 2023 – Present
Architectureresearchfull-stack development
Eldorado is a quantitative research platform for designing, backtesting, and optimizing trading strategies. The project started in 2023 (ConsoleInfluxDb repo): .NET simulation and trading engines, InfluxDB for time-series, SQL Server for relational data, ClickHouse for analytics, and Redis for cache.
The current architecture combines those deterministic engines with a Python AI-agent layer (LangGraph, FastAPI) and local LLMs via Ollama. Machine-learning models (LightGBM, CatBoost) inform the research loop. TimescaleDB is progressively unifying InfluxDB and SQL Server. I structure the system architecture, data pipelines, gRPC contracts, and the React/Vite interface for exploring simulations.
Impact
- •.NET platform since 2023: simulation, live trading, InfluxDB, and SQL Server
- •Simulation and risk engines exposed over gRPC
- •LangGraph AI agents and FastAPI gateway for orchestration
- •Local LLMs via Ollama; LightGBM and CatBoost models in the research loop
- •Time-series evolution: InfluxDB, then TimescaleDB and ClickHouse
- •React/Vite interface for strategy exploration and comparison
Technologies
TypeScriptReactViteBlazorPythonFastAPILangGraph.NET CoreC#gRPCInfluxDBTimescaleDBClickHouseSQL ServerRedisDockerPrometheusGrafanaSeqOllamaLightGBMCatBoostAILLMsAI AgentsMachine LearningGrokGrok BuildDockerGitObservability
Download this mandate