Search

Learn - Docs by LangChain

A guide to LangChain’s Learn section, detailing tutorials and conceptual overviews for building and deploying various AI agents using Managed Deep Agents, Deep Agents SDK, LangChain, and LangGraph, with patterns for multi-agent workflows and customization.

Build It Yourself | Armin Ronacher's Thoughts and Writings

The piece argues for a shift away from heavy dependency use in software development, noting how churn from updates, audits, and platform-layer abstractions burdens Rust and other ecosystems. It advocates writing more self-contained code to reduce maintenance, despite cultural and corporate pressures that favor dependency use.

OPay COO says build for Nigeria's worst days

OPay's COO/CTO argues Nigerian markets require resilience and local adaptation to survive; the company highlights strong NYSE filing, a revenue shift toward lending, and past cash-crunch disruptions as lessons for building durable businesses. Regulation is seen as a design input, not an obstacle.

GitHub - pullboard-dev/pullboard: Autonomous development for real products. Governed agents build from specs you approve, follow the rules you set, and nothing merges until another agent proves it. Your decisions, judgement and taste are the bar that work must clear. Declare your vision and let agents build it. · GitHub

Pullboard systems coordinate multiple autonomous agents to build software from approved specs, with human-approved constraints and independent verification. It emphasizes spec-driven planning, staged verification, lane-based work, and a single human in the loop to declare vision and judge results.

A Template and Usage Guide for Running AI-Driven Development with a 14B-Class Local LLM (Compatible with Cline/Continue/OpenHands and RAG/LLM Wiki) #GenerativeAI - Qiita

Translated title. Original: 14B級ローカルLLMでAI駆動開発を回すためのテンプレートと使用要領（Cline/Continue/OpenHands・RAG/LLM Wiki対応） #生成AI - Qiita

A Bash script generates Cline, Continue, and agent-skill configuration files from shared rules and commands. It checks for unfinished placeholders and applies the selected knowledge-base pattern.

The Asset That AI Models Cannot Build Alone

Translated title. Original: O Ativo Que os Modelos de IA Não Podem Construir Sozinhos

The piece argues that the next AI adoption wave depends on integrating accessible organizational memory around models. It cites Delend’s Second Brain approach—connecting notes, history, and agents—to show how memory, not model size, powers better decisions and reduces misjudgments, highlighting the broader role of memory architecture in AI-enabled enterprises.

GitHub - JCKodel/focus-kit · GitHub

Focus-kit defines a repository delivery workflow guided by a single SETUP.md. It uses a coding agent to generate and apply per-delivery pages, driven by a defined command queue and host-specific prefixes, while preserving project documents and providing feedback via diffs.

lla.ma — build agents, ship them anywhere

lla.ma orchestrates multi-provider build agents with cron-based scheduling, per-model memory, thorough run logging, and per-model budget enforcement to ensure predictable, trackable executions.

Home | AI-orchestrated build playbook

An AI-driven, human-in-the-loop workflow for building a real product uses an orchestrator and temporary subagents to plan, build, test, and review code across multiple repositories. It emphasizes clear design, batch-based execution, audit trails, and decision discipline managed by a human product owner.

AI agent memory. How to build a memory architecture of high quality? / Habr

Translated title. Original: Память ИИ агента. Как собрать архитектуру памяти качественно? / Хабр

The document describes a memory architecture for AI agents that enables cross-agent knowledge sharing using PostgreSQL, pgvector, and a knowledge graph. It outlines memory components, data flow from user messages to memory, and techniques to avoid information loops, with example scales and a focus on sourcing and provenance.