--- title: AI Code Review at Scale: LinkedIn's Multi-Agent Approach description: LinkedIn built an internal multi-agent AI code review platform designed to overcome the limitations of off-the-shelf AI reviewers: single-model blind spots,... canonical: https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: AI Code Review at Scale: LinkedIn's Multi-Agent Approach | daily.dev og:description: LinkedIn built an internal multi-agent AI code review platform designed to overcome the limitations of off-the-shelf AI reviewers: single-model blind spots,... og:url: https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi og:image: https://api.daily.dev/og/posts/7LpKuHKzI.png og:image:alt: AI Code Review at Scale: LinkedIn's Multi-Agent Approach og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # AI Code Review at Scale: LinkedIn's Multi-Agent Approach **[InfoQ](https://daily.dev/sources/infoq)** · 3 min read · 0 upvotes · 0 comments ## Summary LinkedIn built an internal multi-agent AI code review platform designed to overcome the limitations of off-the-shelf AI reviewers: single-model blind spots, weak customization, and lack of operational control. The system runs multiple independent AI reviewers with distinct models, cross-validates their findings, filters low-signal or irrelevant suggestions, and runs on a Kubernetes-based event-driven pipeline that monitors latency, acceptance rates, and provider failures. An acceptance-rate evaluation across 5,230 sampled comments in 1,727 PRs found 63.9% of suggestions were accepted overall, with acceptance varying sharply by category (100% for concurrency bugs, 80% for logic errors, 40.6% for security fixes). The piece also notes Cloudflare's OpenCode-based orchestration system and Databricks' Unity AI Gateway and Omnigent as alternative approaches to AI code review at scale. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What acceptance rate did LinkedIn achieve with its AI code review suggestions? LinkedIn found that 63.9% of AI-generated code review suggestions were accepted by developers, based on an automated pipeline comparing suggestions against the merged codebase across 5,230 sampled review comments in 1,727 pull requests. Acceptance varied widely by category: 100% for concurrency bugs, 80% for logic errors, 58.1% for bug fixes, 43.5% for refactoring changes, and 40.6% for security-related fixes. _Teams weighing AI code review tools can compare real acceptance-rate benchmarks like this one on daily.dev._ ### Why does using a single AI model for code review create blind spots? A single model tends to miss the same class of bugs repeatedly and flag the same low-signal issues, since its blind spots are consistent across every review it performs. LinkedIn addressed this by running multiple independent AI reviewers using distinct models and reasoning approaches, then treating convergence across agents on the same issue as stronger evidence while separately verifying unique findings. _Engineers designing multi-agent review pipelines can track approaches like this via daily.dev._ ### How does LinkedIn's AI code review architecture handle scale and reliability? LinkedIn built the review platform on Kubernetes with an event-driven pipeline, durable queues, and horizontally scaled workers, treating code review as production infrastructure rather than a bolt-on tool. This design enables monitoring of latency, acceptance and completion rates, and provider failures, giving operational control that off-the-shelf AI reviewers lack. _daily.dev helps engineers building production-grade AI infrastructure follow architectures like this one._ ## Similar posts on daily.dev - [Orchestrating AI Code Review at scale](https://daily.dev/posts/orchestrating-ai-code-review-at-scale-re4wh2np4) · Cloudflare · 0 upvotes · 0 comments - [Agentic Code Review](https://daily.dev/posts/agentic-code-review-0ue8fcsa1) · Addy Osmani · 147 upvotes · 14 comments - [How DoorDash built an AI code reviewer engineers actually listen to](https://daily.dev/posts/how-doordash-built-an-ai-code-reviewer-engineers-actually-listen-to-psc8auyls) · Doordash · 0 upvotes · 0 comments - [AI Is Moving up the Software Lifecycle: From Code Review to PRD Governance](https://daily.dev/posts/ai-is-moving-up-the-software-lifecycle-from-code-review-to-prd-governance-pkxs7diem) · InfoQ · 0 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#kubernetes](https://daily.dev/tags/kubernetes), [#ai-agents](https://daily.dev/tags/ai-agents), [#code-review](https://daily.dev/tags/code-review), [#linkedin](https://daily.dev/tags/linkedin) [View this post on daily.dev](https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"AI Code Review at Scale: LinkedIn's Multi-Agent Approach","url":"https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi"},"datePublished":"2026-08-22T09:02:51.878Z","dateModified":"2026-09-13T20:59:22.575Z","description":"LinkedIn built an internal multi-agent AI code review platform designed to overcome the limitations of off-the-shelf AI reviewers: single-model blind spots,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/ef8fa3a710c2be32a29fe1d2a72040b8?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/ef8fa3a710c2be32a29fe1d2a72040b8?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"InfoQ","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"InfoQ","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/afc3bced3e1e4b188dd9127017a60e0c","url":"https://daily.dev/sources/infoq"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,kubernetes,ai-agents,code-review,linkedin","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"InfoQ","item":"https://daily.dev/sources/infoq"},{"@type":"ListItem","position":3,"name":"AI Code Review at Scale: LinkedIn's Multi-Agent Approach"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi#faq","mainEntity":[{"@type":"Question","name":"What acceptance rate did LinkedIn achieve with its AI code review suggestions?","acceptedAnswer":{"@type":"Answer","text":"LinkedIn found that 63.9% of AI-generated code review suggestions were accepted by developers, based on an automated pipeline comparing suggestions against the merged codebase across 5,230 sampled review comments in 1,727 pull requests. Acceptance varied widely by category: 100% for concurrency bugs, 80% for logic errors, 58.1% for bug fixes, 43.5% for refactoring changes, and 40.6% for security-related fixes. Teams weighing AI code review tools can compare real acceptance-rate benchmarks like this one on daily.dev."}},{"@type":"Question","name":"Why does using a single AI model for code review create blind spots?","acceptedAnswer":{"@type":"Answer","text":"A single model tends to miss the same class of bugs repeatedly and flag the same low-signal issues, since its blind spots are consistent across every review it performs. LinkedIn addressed this by running multiple independent AI reviewers using distinct models and reasoning approaches, then treating convergence across agents on the same issue as stronger evidence while separately verifying unique findings. Engineers designing multi-agent review pipelines can track approaches like this via daily.dev."}},{"@type":"Question","name":"How does LinkedIn's AI code review architecture handle scale and reliability?","acceptedAnswer":{"@type":"Answer","text":"LinkedIn built the review platform on Kubernetes with an event-driven pipeline, durable queues, and horizontally scaled workers, treating code review as production infrastructure rather than a bolt-on tool. This design enables monitoring of latency, acceptance and completion rates, and provider failures, giving operational control that off-the-shelf AI reviewers lack. daily.dev helps engineers building production-grade AI infrastructure follow architectures like this one."}}]} ```