Presentation: Context Engineering at LinkedIn: How We Built an Organizational Context Layer for AI Agents with MCP
Ajay Prakash discusses how LinkedIn overcomes AI agent limitations in large codebases. He explains Contextual Agent Playbooks and Tools - built on Model Context Protocol (MCP) - which serves procedural memory, code search, and runbooks directly to coding agents. Prakash shares architectural details and operational guardrails that deliver a 20% productivity boost with zero loss in reliability. By…
Ajay Prakash, a software engineer at LinkedIn, shares how the company built an organizational context layer for AI agents using the Model Context Protocol (MCP). This system, called Contextual Agent Playbooks and Tools, helps coding agents understand the internal context of LinkedIn's large, mature codebase. Despite the excitement around AI coding assistants and "vibe coding," early attempts proved ineffective due to the lack of contextual understanding.
Prakash explains that when an engineer faces a critical issue, such as a latency spike affecting millions of users, a coding agent can quickly fetch necessary instructions, identify the root cause, and take action. This process, which would normally take hours, is accomplished in just minutes with the help of the new system. The Contextual Agent Playbooks and Tools consist of over 600 workflows and thousands of tools that automate workflows and boost productivity for LinkedIn teams.
Prakash concludes by highlighting the system's success and the lessons learned during its development.
Written by urgent.news from InfoQ's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
