AlphaForge – Container View (C2)

Source: Notion | Last edited: 2025-11-21 | ID: 2b22d2dc-3ef...

---

Overview

This view describes the main containers inside AlphaForge / QuantOS and how they interact with users, data providers, and execution venues.

You can optionally turn each container into a toggle block in Notion to keep the page compact.

---

High-Level Containers

API and Gateway

Entry point for users and AI agents.

• Exposes HTTP / gRPC interfaces
• Handles authentication and authorization
• Routes requests to internal services:
  • DSL and Compiler Service
  • Orchestrator Service
  • Experiment and Metric Store

---

DSL and Compiler Service

Responsible for understanding and validating the strategy DSL.

• Parses and validates strategy DSL
• Produces:
  • IR (Intermediate Representation)
  • Canonical DAGs
  • Fingerprints
  • Vector embeddings
• Provides APIs for:
  • Compile
  • Lint
  • Static analysis

---

Orchestrator Service

Brains of backtests, simulations, paper trading, and live runs.

• Manages runs as jobs
• Schedules workloads on compute resources
• Tracks:
  • Run state
  • Artifacts
  • Logs
  • Lifecycle events (created, running, completed, failed, promoted, etc.)

---

Data and Feature Service

Unified layer for historical & live data plus derived features.

• Provides consistent access to:
  • Historical data (e.g. via ClickHouse, object storage)
  • Live market feeds
• Handles:
  • Ingestion and ETL
  • Normalization across exchanges and venues
  • Schema management for market and on-chain data
• Generates and serves:
  • Feature panels
  • Factor panels

---

Execution Gateway

Bridge between AlphaForge and execution engines / venues.

• Connects to:
  • Exchanges
  • Brokers
  • External execution engines (e.g. Nautilus, custom engines)
• Translates:
  • Abstract orders and strategy intents → venue-specific API calls
• Receives and streams back to Orchestrator:
  • Fills
  • Positions
  • Account state

---

Plugin and Capability Pack Registry

Registry for plugins contributed by capability packs.

• Indexes and manages plugins from:
  • Mid-frequency packs
  • HFT packs
  • On-chain / DeFi packs
• Stores metadata about:
  • Data sources
  • Features
  • Signals
  • Execution adapters
• Exposes discovery APIs used by:
  • DSL and Compiler Service
  • Orchestrator Service

---

Experiment and Metric Store

Source of truth for experiments and results.

• Stores:
  • Experiment definitions
  • Runs
  • Metrics and curves
  • Artifacts and metadata
• Supports search and querying by:
  • Fingerprint
  • Tags
  • Performance metrics
  • Other strategy metadata
• Powers:
  • Dashboards
  • Notebooks
  • Agent queries (for AI research agents)

---

Monitoring and Observability Stack

Operational visibility for the whole platform.

• Collects:
  • Metrics
  • Logs
  • Traces
• Powers:
  • Dashboards for platform health
  • Alerting for errors, latency, resource issues

---

Identity and Access Management (IAM)

Manages who can do what inside AlphaForge.

• Manages:
  • Users and teams
  • Roles and permissions
• Enforces:
  • Access control on DSL, runs, data, and execution
• Can integrate with external identity providers (SSO, corporate IdP) if needed

---

Container Diagram (Optional)

In Notion, add a Code block, set language to mermaid, and paste:

flowchart LR
subgraph Client
    R[Quant Researcher]
    A[AI Agent]
    PM[Portfolio Manager]
end

subgraph AlphaForge
    APIGW[API and Gateway]
    COMP[DSL and Compiler Service]
    ORCH[Orchestrator Service]
    DATA[Data and Feature Service]
    EXEC[Execution Gateway]
    PLUG[Plugin and Capability Pack Registry]
    EXP[Experiment and Metric Store]
end

subgraph External
    MKT[Market and On-Chain Data Sources]
    VENUES[Exchanges, Brokers, Execution Engines]
    STORE[Object Storage and Data Warehouse]
end

R --> APIGW
A --> APIGW
PM --> APIGW

APIGW --> COMP
APIGW --> ORCH
APIGW --> EXP

COMP --> PLUG
ORCH --> DATA
ORCH --> EXEC
ORCH --> EXP
DATA --> STORE

MKT --> DATA
EXEC --> VENUES