Full Workshop: Build Your Own Deep Research Agents - Louis-François Bouchard, Paul Iusztin, Samridhi

2026-04-20 · Source: AI Engineer · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, extended

Summary

Towards AI presented a workshop on automating technical content creation, addressing the prevalence of generic, "AI slop" content. The core problem identified is the high cost and time required for senior AI engineers to produce quality technical articles, videos, and training materials. Their solution involves a two-part automated system: a deep research agent and a writing workflow. The deep research agent autonomously plans, searches the web, analyzes YouTube videos and GitHub content, and synthesizes information into a research markdown file. The writing workflow then takes this research, along with user-defined guidelines and writing profiles, to generate, review, and edit high-quality LinkedIn posts, aiming to avoid AI-generated characteristics and ensure human-like tone. The system emphasizes a human-in-the-loop approach, particularly for the creative and relational aspects of writing, and utilizes observability tools like OPIC for monitoring and AI evals for quality assurance and regression testing.

Key takeaway

For AI Engineers building content generation systems, you should design distinct, specialized agents for research and writing tasks. The research agent needs flexibility for exploration, while the writing workflow requires strict constraints and an evaluator-optimizer loop to ensure quality and avoid "AI slop." This approach minimizes costs and improves content reliability, but remember to integrate human oversight for nuanced, relatable output.

Key insights

Automating technical content production requires distinct research and writing systems, balancing flexibility with strict constraints.

Principles

• Prioritize simplest solutions first.
• Separate exploratory research from deterministic writing.
• Maintain human-in-the-loop for creative content.

Method

The proposed method involves a deep research agent for data gathering and a separate writing workflow for content generation, review, and editing, using structured inputs and an evaluator-optimizer loop.

In practice

• Use structured guidelines for LLM content generation.
• Implement writing profiles to control tone and style.
• Employ few-shot examples for improved LLM performance.

Topics

• Deep Research Agents
• AI Content Automation
• Technical Writing Workflow
• LLM Judge
• Observability

Best for: AI Engineer, MLOps Engineer, AI Student

Related on AIssential

• Google Deep Research Max: Build Autonomous AI Research Agents in Minutes — Deep Research Max is an autonomous AI agent for comprehensive, cited research and native data visualization via a single API call.
• Fantastic Scientific Agents and How to Build Them: AgentBuild for Rietveld Refinement — AgentBuild constructs LLM agents from scientist-authored contracts, preserving expert judgment and enabling durable, re-tunable scientific workflows.
• The AI Scientist: Towards Fully Automated AI Research, Now Published in Nature — AI agents can now autonomously conduct machine learning research, generate papers, and pass human peer review.
• The 12-Hour Novel Factory: What do we want ‘a book’ to mean in a world where text is abundant—and what governance, provenance, and value signals will we enforce to protect that meaning? — Orchestrating AI agent teams transforms creative writing into an industrialized, quality-controlled workflow.
• Google’s new Deep Research and Deep Research Max agents can search the web and your private data — Google's new Deep Research agents fuse web and private data, generating visual reports for enterprise workflows.
• DeepMind’s New AI Just Changed Science Forever — DeepMind's Aletheia AI can autonomously conduct research and generate novel scientific content, pushing the boundaries of AI capabilities.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.