Trying to write a non-ai article about ai automating things (specifically quant research)

2026-04-11 · Source: LLM on Medium · Field: Finance & Economics — Capital Markets & Investment Management, FinTech & Digital Financial Services · Depth: Intermediate, medium

Summary

A Product leader and quantitative researcher details a production-grade system for backtesting, execution, and research in algorithmic trading, leveraging Large Language Models (LLMs). The system integrates existing workflows with AI tools to automate and scale quant research, addressing challenges like feature engineering and exploring new strategies. Key components include Claude for orchestration, Obsidian for qualitative research and idea linking, and Linear for structured workflow management across qualitative research, engineering, quantitative research, and ideation. The author emphasizes the importance of defining "definition of done" for each workflow state to ensure agents maintain focus and deliver verifiable outputs, significantly enhancing the ability to iterate and improve research processes.

Key takeaway

For AI Architects designing automated research systems, you should focus on establishing clear, verifiable workflow states for each agent. This approach, akin to a "definition of done" for every step, prevents shortcuts and ensures consistent, high-quality outputs, allowing your teams to iterate and improve processes effectively. Consider integrating tools like Obsidian for context management and Linear for workflow orchestration to maximize efficiency.

Key insights

Integrating LLMs into structured workflows significantly scales quantitative research and automates complex tasks.

Principles

• Define "definition of done" for every workflow state.
• Modular context improves agent parsing and performance.

Method

Utilize an orchestrator LLM (e.g., Claude) to manage agent teams, integrate a note-taking app (Obsidian) for linking qualitative and quantitative research, and use a project management tool (Linear) to structure and A/B test workflows.

In practice

• Use Obsidian to link qualitative research to experiments.
• Implement Kanban-like boards for agent task assignment.
• Build dashboards with Mermaid diagrams for results visualization.

Topics

• Quant Research Automation
• LLM-driven Workflows
• Agent Teams
• Obsidian Knowledge Graph
• Algorithmic Trading Strategies

Best for: AI Architect, AI Engineer, Machine Learning Engineer, Director of AI/ML

Related on AIssential

• How I'm Using AI Today — A structured, principled approach to AI agent management mitigates LLM context limitations and enhances productivity.
• Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI — AI agents are fundamentally transforming software development and research by enabling extensive task delegation and automation.
• Another try at LLM workflow automation — Effective LLM automation requires extensive scaffolding and iterative tooling, not just smarter models.
• Local AI: The 5-Hour Knowledge Graph — Effective agentic AI requires meticulously defined workflows, not just capable technology.
• From Research Question to Scientific Workflow: Leveraging Agentic AI for Science Automation — An agentic AI architecture automates scientific workflow generation from natural language, ensuring deterministic execution via layered design.
• TAI #207: Claude Opus 4.8 Is Better, but Dynamic Workflows Are the Bigger Story — Dynamic Workflows, where LLMs orchestrate multi-agent tasks, represent a significant shift towards scalable, complex problem-solving, despite increased token costs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by LLM on Medium.