--- title: Developers should be aware of these two AI engineering... description: Two emerging engineering practices for building reliable AI agentic systems are introduced: Graph Engineering, which breaks work into smaller connected tasks... canonical: https://daily.dev/posts/developers-should-be-aware-of-these-two-ai-engineering-approaches-lgfvrktej twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Developers should be aware of these two AI engineering approaches | daily.dev og:description: Two emerging engineering practices for building reliable AI agentic systems are introduced: Graph Engineering, which breaks work into smaller connected tasks... og:url: https://daily.dev/posts/developers-should-be-aware-of-these-two-ai-engineering-approaches-lgfvrktej og:image: https://api.daily.dev/og/posts/LgFVrktej.png og:image:alt: Developers should be aware of these two AI engineering approaches 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. # Developers should be aware of these two AI engineering approaches **[Pavan Belagatti](https://daily.dev/sources/dfa0u7-fs)** · [@johhnypav](https://daily.dev/johhnypav) · 2 min read · 136 upvotes · 20 comments ## Summary Two emerging engineering practices for building reliable AI agentic systems are introduced: Graph Engineering, which breaks work into smaller connected tasks with explicit inputs/outputs instead of one massive prompt, and Harness Engineering, which gives AI coding agents the broader context (service ownership, dependencies, rules, permissions, verification methods) needed beyond mere code generation. Links to video guides covering a payment bug fix example and a scorecard-based deployment gate are included. ## Content As a developer, there are two engineering practices you should be aware of - Harness Engineering & Graph Engineering, to build reliable agentic systems. **Graph Engineering** is becoming one of the most useful ways to build reliable AI-native engineering systems. Instead of putting everything into one massive prompt and hoping an AI model handles every detail correctly, I design the work as a connected system of smaller, explicit tasks. Each task has a purpose, a defined input, an output, and a route to the next step. That is the core idea behind Graph Engineering. It gives AI agents structure, makes their decisions easier to inspect, and lets independent work happen in parallel. Similarly, Harness Engineering is becoming essential because AI coding agents are getting very good at generating code, but code generation is only one small part of real software engineering. A useful agent must also understand the service it is working on, the team that owns it, its dependencies, the engineering rules it must follow, what it is allowed to change, and how to verify that its work is correct. That complete environment is what makes an AI agent reliable. In this guide, I will explain Harness Engineering in a simple way, then walk through two practical implementations: a controlled payment bug fix and a scorecard-based deployment gate. Well, I would like to share the guides around these two engineering approaches. Here is my video guide on harness engineering: [https://youtu.be/vAHnGUWPPNY](https://youtu.be/vAHnGUWPPNY?si=kMliPXcr6L3Jp6N5) Here is my video guide on graph engineering: [https://youtu.be/g3C7mudBURA](https://youtu.be/g3C7mudBURA?si=xrF8qOMFNjovsV7j) ## Community discussion Top comments from developers on daily.dev. **@andrewduncan** · 8 upvotes > interesting article! Small edit @johhnypav your thumbnail has the descriptions backwards. the description for graph engineering appears under harness engineering, and the description for harness engineering appears under graph engineering. > > Graph engineering has fascinating potential **@harshitmanepally** · 3 upvotes > A great article. Helpful for beginners. **@swarnasharmin** · 3 upvotes > Graph architecture paired with strict harness rules is exactly how we move from flaky LLM demos to production-grade agents. **@ahmetozel** · 2 upvotes > The parallel-work point is where the graph needs a contract for joining results. If two independent tasks revise the same artifact, a successful result from each does not establish that their combined output is consistent. The merge itself becomes a step with inputs and a failure mode. > > I would make ownership of each output explicit, then test one branch finishing after another has invalidated its assumptions. The harness can enforce access and execution limits, while the graph decides whether the late result is still usable or must be recomputed. That distinction is useful even when every... **@diyweb** · 1 upvotes > Kiran's note on smaller tasks for debugging matches the graph side well. A harness that sets tools, sandbox and guardrails still fails if the workflow is one opaque run. The payment bug-fix and scorecard gate examples make the pair concrete without needing the video. ## Similar posts on daily.dev - [The rise of agentic engineering](https://daily.dev/posts/the-rise-of-agentic-engineering-arnnwynck) · Data Science Central · 3 upvotes · 0 comments - [Graph Engineering for AI Agents: From Prompts and Loops to Workflows](https://daily.dev/posts/graph-engineering-for-ai-agents-from-prompts-and-loops-to-workflows-dblplaofw) · Towards Data Science · 0 upvotes · 0 comments - [Harness Engineering](https://daily.dev/posts/harness-engineering-l1sahcwor) · Martin Fowler · 4 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#architecture](https://daily.dev/tags/architecture), [#prompt-engineering](https://daily.dev/tags/prompt-engineering) [View this post on daily.dev](https://daily.dev/posts/developers-should-be-aware-of-these-two-ai-engineering-approaches-lgfvrktej) ```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":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/developers-should-be-aware-of-these-two-ai-engineering-approaches-lgfvrktej","headline":"Developers should be aware of these two AI engineering approaches","text":"Two emerging engineering practices for building reliable AI agentic systems are introduced: Graph Engineering, which breaks work into smaller connected tasks with explicit inputs/outputs instead of one massive prompt, and Harness Engineering, which gives AI coding agents the broader context (service ownership, dependencies, rules, permissions, verification methods) needed beyond mere code generation. Links to video guides covering a payment bug fix example and a scorecard-based deployment gate are included.","url":"https://daily.dev/posts/developers-should-be-aware-of-these-two-ai-engineering-approaches-lgfvrktej","datePublished":"2026-10-11T16:58:28.770Z","dateModified":"2026-10-12T04:45:09.223Z","author":{"@type":"Person","name":"Pavan Belagatti","url":"https://daily.dev/johhnypav","image":"https://media.daily.dev/image/upload/s--r9dshbly--/f_auto/v1776930121/avatars/avatar_DFA0U7-fS?_a=BAMAMiWQ0","description":"Developer Evangelist","interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"EndorseAction"},"userInteractionCount":4680}},"image":"https://media.daily.dev/image/upload/s--bKMdkuJE--/f_auto/v1791780308/posts/LgFVrktej?_a=BAMAMicg0","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":136},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":20}],"comment":[{"@type":"Comment","text":"interesting article! Small edit @johhnypav your thumbnail has the descriptions backwards. the description for graph engineering appears under harness engineering, and the description for harness engineering appears under graph engineering.\nGraph engineering has fascinating potential","datePublished":"2026-10-12T02:19:07.353Z","url":"https://daily.dev/posts/LgFVrktej#c-RtMV71QT6","author":{"@type":"Person","name":"Andrew Duncan","url":"https://daily.dev/andrewduncan","image":"https://lh3.googleusercontent.com/a/ACg8ocIOpyCFieHnXPAc199x8b2A7uhSGmRANbdNykfN9iOaYPIG6vU=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":8}},{"@type":"Comment","text":"A great article. Helpful for beginners.","datePublished":"2026-10-12T05:36:14.922Z","url":"https://daily.dev/posts/LgFVrktej#c-aeflBehyu","author":{"@type":"Person","name":"HARSHIT MANEPALLY","url":"https://daily.dev/harshitmanepally","image":"https://media.daily.dev/image/upload/s--O0TOmw4y--/f_auto/v1715772965/public/noProfile"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":3}},{"@type":"Comment","text":"Graph architecture paired with strict harness rules is exactly how we move from flaky LLM demos to production-grade agents.","datePublished":"2026-10-12T05:51:56.891Z","url":"https://daily.dev/posts/LgFVrktej#c-w84UkgIml","author":{"@type":"Person","name":"Swarna Sharmin","url":"https://daily.dev/swarnasharmin","image":"https://media.daily.dev/image/upload/s--joHEq_jD--/f_auto/v1791022797/avatars/avatar_ZL6bebyCVkszL2cr0g8ab?_a=BAMAMicg0"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":3}},{"@type":"Comment","text":"The parallel-work point is where the graph needs a contract for joining results. If two independent tasks revise the same artifact, a successful result from each does not establish that their combined output is consistent. The merge itself becomes a step with inputs and a failure mode.\nI would make ownership of each output explicit, then test one branch finishing after another has invalidated its assumptions. The harness can enforce access and execution limits, while the graph decides whether the late result is still usable or must be recomputed. That distinction is useful even when every individual node reports success.","datePublished":"2026-10-15T20:55:05.176Z","url":"https://daily.dev/posts/LgFVrktej#c-oidbg8jKq","author":{"@type":"Person","name":"Ahmet Özel","url":"https://daily.dev/ahmetozel","image":"https://avatars.githubusercontent.com/u/70992231?v=4"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2}},{"@type":"Comment","text":"Kiran’s note on smaller tasks for debugging matches the graph side well. A harness that sets tools, sandbox and guardrails still fails if the workflow is one opaque run. The payment bug-fix and scorecard gate examples make the pair concrete without needing the video.","datePublished":"2026-10-14T15:32:29.132Z","url":"https://daily.dev/posts/LgFVrktej#c-PzPfnp2Pl","author":{"@type":"Person","name":"DiyWeb","url":"https://daily.dev/diyweb","image":"https://media.daily.dev/image/upload/s--dgS6jDeN--/f_auto/v1790879420/avatars/avatar_774zDwMaRTNS1fv5OKStX?_a=BAMAMicg0"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/dfa0u7-fs","name":"Pavan Belagatti"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Pavan Belagatti","item":"https://daily.dev/sources/dfa0u7-fs"},{"@type":"ListItem","position":3,"name":"Developers should be aware of these two AI engineering approaches"}]} ```