AI Trading Agent Architect (Backend & Evaluation)

3Commas; España; On-site; EUR 65,000 - 90,000; Full time

Job summary

3Commas is hiring for an Applied AI backend/ML engineer to own the AI research engine behind QuantPilot's agents. You will focus on research, strategy generation, backtesting, and the agent platform, delivering statistically sound results with robust production-grade systems.

You will drive end-to-end architecture, ensure contracts between engine and product, and build the evaluation layer with regression detection. Strong Python, Kubernetes, Go and TS skills are essential.

Qualifications

• 5+ years of backend or ML engineering experience.
• At least 1 year building LLM-based systems in production.
• Track record of eval-driven development with measurable quality metrics.

Responsibilities

• Own the agent architecture end to end: orchestration, tool loops, context management, sandboxes and model routing.
• Own the contracts between the engine and the product, including frontend rendering of agent events.
• Build and own the evaluation layer: offline evals, backtest-based quality metrics and regression detection.
• Bring statistical rigor to strategy evaluation: overfitting checks, robustness, walk-forward and out-of-sample validation.
• Build data foundations with provenance and a scalable model of market entities; ensure observability for non-deterministic failures.
• Shape product direction through RFCs and ADRs; mentor engineers moving into LLM systems.

Skills

Python

Kubernetes

Go

TypeScript

LLM systems

English B2+

Eval-driven development

Tools

MCP protocol

Job description

3Commas is hiring for an Applied AI backend/ML engineer to own the AI research engine behind QuantPilot's agents. You will focus on research, strategy generation, backtesting, and the agent platform, delivering statistically sound results with robust production-grade systems.

You will drive end-to-end architecture, ensure contracts between engine and product, and build the evaluation layer with regression detection. Strong Python, Kubernetes, Go and TS skills are essential.