{"id":31431,"date":"2026-05-24T16:44:48","date_gmt":"2026-05-24T13:44:48","guid":{"rendered":"https:\/\/immediatech.net\/personal\/?p=31431"},"modified":"2026-06-03T16:59:46","modified_gmt":"2026-06-03T13:59:46","slug":"from-prompts-to-state-machines-designing-cyclic-multi-agent-workflows-with-langgraph","status":"publish","type":"post","link":"https:\/\/immediatech.net\/personal\/from-prompts-to-state-machines-designing-cyclic-multi-agent-workflows-with-langgraph\/","title":{"rendered":"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"31431\" class=\"elementor elementor-31431\" data-elementor-post-type=\"post\">\n\t\t\t\t<div class=\"elementor-element elementor-element-83e7931 e-flex e-con-boxed e-con e-parent\" data-id=\"83e7931\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-ad803e4 elementor-widget elementor-widget-text-editor\" data-id=\"ad803e4\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t<h2 data-path-to-node=\"4\">Introduction<\/h2><p data-path-to-node=\"5\">Early generative AI implementations relied heavily on linear prompting: a user submits a query, an LLM processes it, and a static output is generated. While this works for basic generation, enterprise-level automation demands systems that can reason, self-correct, and execute complex, multi-step processes autonomously.<\/p><p data-path-to-node=\"6\">To build robust, reliable AI features, architecture must move beyond simple chains. The frontier of production-ready GenAI lies in <b data-path-to-node=\"6\" data-index-in-node=\"131\">cyclic multi-agent workflows<\/b> managed as deterministic state machines. This is where <b data-path-to-node=\"6\" data-index-in-node=\"215\">LangGraph<\/b> becomes a critical component of the enterprise AI stack.<\/p><h2 data-path-to-node=\"8\">The Limitations of Linear LLM Chains<\/h2><p data-path-to-node=\"9\">Standard LLM orchestration frameworks often rely on Directed Acyclic Graphs (DAGs). In a DAG, data flows strictly forward:<\/p><div data-path-to-node=\"10\"><div class=\"math-block\" data-math=\"\\text{Input} \\longrightarrow \\text{Agent A} \\longrightarrow \\text{Agent B} \\longrightarrow \\text{Output}\">$$\\text{Input} \\longrightarrow \\text{Agent A} \\longrightarrow \\text{Agent B} \\longrightarrow \\text{Output}$$<\/div><\/div><p data-path-to-node=\"11\">While efficient for predictable pipelines, linear architectures break down when faced with complex enterprise requirements that demand loops and iterative refinement:<\/p><ul data-path-to-node=\"12\"><li><p data-path-to-node=\"12,0,0\"><b data-path-to-node=\"12,0,0\" data-index-in-node=\"0\">No Self-Correction:<\/b> If Agent B identifies an error in Agent A&#8217;s output, a linear chain cannot route the data backward for a rewrite.<\/p><\/li><li><p data-path-to-node=\"12,1,0\"><b data-path-to-node=\"12,1,0\" data-index-in-node=\"0\">Lack of State Persistence:<\/b> Complex, multi-turn business logic requires a centralized, mutable state that agents can read from and write to concurrently.<\/p><\/li><li><p data-path-to-node=\"12,2,0\"><b data-path-to-node=\"12,2,0\" data-index-in-node=\"0\">Deterministic Friction:<\/b> Purely autonomous agent swarms without strict structural boundaries often suffer from infinite loops or prompt drift, wasting costly API tokens.<\/p><\/li><\/ul><h2 data-path-to-node=\"14\">Why State Machines? Enter LangGraph<\/h2><p data-path-to-node=\"15\">LangGraph addresses these limitations by allowing developers to model agent interactions as a <b data-path-to-node=\"15\" data-index-in-node=\"94\">State Machine<\/b>. By defining explicit nodes, edges, and a centralized state, you gain total control over agent collaboration while allowing the system to run cyclic loops safely.<\/p><h3 data-path-to-node=\"16\">The Anatomy of a LangGraph Workflow<\/h3><ol start=\"1\" data-path-to-node=\"17\"><li><p data-path-to-node=\"17,0,0\"><b data-path-to-node=\"17,0,0\" data-index-in-node=\"0\">State:<\/b> A shared, mutable data structure (e.g., a Python <code data-path-to-node=\"17,0,0\" data-index-in-node=\"56\">TypedDict<\/code> or Pydantic model) that maintains the ground truth across the entire lifecycle of the execution.<\/p><\/li><li><p data-path-to-node=\"17,1,0\"><b data-path-to-node=\"17,1,0\" data-index-in-node=\"0\">Nodes:<\/b> Isolated functional blocks or specialized agents (e.g., a Python worker using <b data-path-to-node=\"17,1,0\" data-index-in-node=\"85\">Playwright<\/b> for web scraping, or an LLM grader evaluating code quality).<\/p><\/li><li><p data-path-to-node=\"17,2,0\"><b data-path-to-node=\"17,2,0\" data-index-in-node=\"0\">Edges:<\/b> Relational pathways that determine routing. LangGraph introduces <i data-path-to-node=\"17,2,0\" data-index-in-node=\"72\">conditional edges<\/i>, allowing the system to decide the next execution path dynamically based on the current state.<\/p><\/li><\/ol><div class=\"code-block ng-tns-c3025443569-142 ng-animate-disabled ng-trigger ng-trigger-codeBlockRevealAnimation\" data-hveid=\"0\" data-ved=\"0CAAQhtANahgKEwim0ayy0eqUAxUAAAAAHQAAAAAQ7As\"><div class=\"formatted-code-block-internal-container ng-tns-c3025443569-142\"><div class=\"animated-opacity ng-tns-c3025443569-142\"><pre class=\"ng-tns-c3025443569-142\"><code class=\"code-container formatted ng-tns-c3025443569-142 no-decoration-radius\" role=\"text\" data-test-id=\"code-content\">                  +-----------------------+\n                  |      User Input       |\n                  +-----------+-----------+\n                              |\n                              v\n                  +-----------+-----------+\n                  |   Node 1: Researcher  |&lt;----+\n                  +-----------+-----------+     |\n                              |                 | (Cyclic Loop)\n                              v                 |\n                  +-----------+-----------+     |\n                  |    Node 2: Critic     |-----+\n                  +-----------+-----------+\n                              |\n                     (Validation Passed)\n                              v\n                  +-----------+-----------+\n                  |   Final Output\/Sync   |\n                  +-----------------------+\n<\/code><\/pre><\/div><\/div><\/div><h2 data-path-to-node=\"20\">Blueprint: Designing a Cyclic Multi-Agent System<\/h2><p data-path-to-node=\"21\">Consider a production use case: an automated pipeline that ingests raw business data, evaluates it against compliance rules, and synchronizes the validated output directly into an enterprise system like <b data-path-to-node=\"21\" data-index-in-node=\"203\">Salesforce<\/b>.<\/p><p data-path-to-node=\"22\">Instead of trusting one prompt to handle the entire operation, we divide the responsibility into a specialized multi-agent graph:<\/p><h3 data-path-to-node=\"23\">Step 1: The Researcher Node<\/h3><p data-path-to-node=\"24\">Equipped with automation tools like <b data-path-to-node=\"24\" data-index-in-node=\"36\">Playwright<\/b>, this agent scrapes, parses, and structures high-volume raw data from target sources, updating the graph&#8217;s <code data-path-to-node=\"24\" data-index-in-node=\"154\">State<\/code> with a draft payload.<\/p><h3 data-path-to-node=\"25\">Step 2: The Critic \/ Validator Node<\/h3><p data-path-to-node=\"26\">A specialized LLM node evaluates the draft payload against a strict schema (such as GDPR alignment or internal formatting guidelines).<\/p><h3 data-path-to-node=\"27\">Step 3: The Conditional Routing Edge<\/h3><p data-path-to-node=\"28\">This is where the cyclic nature occurs. The conditional edge checks the Validator&#8217;s score:<\/p><ul data-path-to-node=\"29\"><li><p data-path-to-node=\"29,0,0\"><b data-path-to-node=\"29,0,0\" data-index-in-node=\"0\">If Validation Fails:<\/b> The edge routes the data <b data-path-to-node=\"29,0,0\" data-index-in-node=\"46\">back<\/b> to the Researcher node, appending the error logs to the state. The Researcher reads the feedback, refines the script execution, and submits a v2.<\/p><\/li><li><p data-path-to-node=\"29,1,0\"><b data-path-to-node=\"29,1,0\" data-index-in-node=\"0\">If Validation Passes:<\/b> The data routes forward to an execution node that triggers the API call to finalize the workflow.<\/p><\/li><\/ul><h2 data-path-to-node=\"31\">Business Impact: Slashing OpEx and Scaling Reliability<\/h2><p data-path-to-node=\"32\">Shifting from fragile prompts to resilient state machines drastically changes operational efficiency:<\/p><ul data-path-to-node=\"33\"><li><p data-path-to-node=\"33,0,0\"><b data-path-to-node=\"33,0,0\" data-index-in-node=\"0\">Predictable Token Consumption:<\/b> Guardrails prevent runaway executions, keeping infrastructure costs highly optimized.<\/p><\/li><li><p data-path-to-node=\"33,1,0\"><b data-path-to-node=\"33,1,0\" data-index-in-node=\"0\">Zero-Downtime Data Integration:<\/b> By isolating agent responsibilities into individual nodes, debugging becomes modular. If a data sync fails, only the execution node retries, leaving the generated context intact.<\/p><\/li><li><p data-path-to-node=\"33,2,0\"><b data-path-to-node=\"33,2,0\" data-index-in-node=\"0\">True High-Agency Automation:<\/b> Systems operate with a 40% reduction in manual operational overhead, translating complex, volatile web logic into clean enterprise profit.<\/p><\/li><\/ul><h2 data-path-to-node=\"35\">Conclusion<\/h2><p data-path-to-node=\"36\">Prompt engineering is a baseline skill; <b data-path-to-node=\"36\" data-index-in-node=\"40\">Graph Architecture<\/b> is the actual future of enterprise AI automation. Tools like LangGraph enable engineers to construct deterministic, resilient, and cyclic agent networks capable of handling complex business processes without human intervention.<\/p><p data-path-to-node=\"37\">To future-proof your digital infrastructure, stop writing longer prompts. Start building better state machines.<\/p><h3 data-path-to-node=\"39\">\u00a0<\/h3>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Move beyond linear LLM chains. Learn how to design resilient, cyclic multi-agent AI architectures and state machines using LangGraph and Python.<\/p>\n","protected":false},"author":1,"featured_media":31300,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[1],"tags":[],"class_list":["post-31431","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-insights"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.7 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph - ANTON OSHMIAN<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph - ANTON OSHMIAN\" \/>\n<meta property=\"og:description\" content=\"Move beyond linear LLM chains. Learn how to design resilient, cyclic multi-agent AI architectures and state machines using LangGraph and Python.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\" \/>\n<meta property=\"og:site_name\" content=\"ANTON OSHMIAN\" \/>\n<meta property=\"article:published_time\" content=\"2026-05-24T13:44:48+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-06-03T13:59:46+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"2240\" \/>\n\t<meta property=\"og:image:height\" content=\"1260\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"antoshby\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"antoshby\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\"},\"author\":{\"name\":\"antoshby\",\"@id\":\"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36\"},\"headline\":\"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph\",\"datePublished\":\"2026-05-24T13:44:48+00:00\",\"dateModified\":\"2026-06-03T13:59:46+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\"},\"wordCount\":663,\"publisher\":{\"@id\":\"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36\"},\"image\":{\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg\",\"articleSection\":[\"insights\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\",\"url\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\",\"name\":\"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph - ANTON OSHMIAN\",\"isPartOf\":{\"@id\":\"https:\/\/immediatech.net\/personal\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg\",\"datePublished\":\"2026-05-24T13:44:48+00:00\",\"dateModified\":\"2026-06-03T13:59:46+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/immediatech.net\/personal\/transition-manifesto\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage\",\"url\":\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg\",\"contentUrl\":\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg\",\"width\":2240,\"height\":1260,\"caption\":\"I architect AI automation, web systems & data pipelines that reduce operational costs.\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/immediatech.net\/personal\/transition-manifesto\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Personal\",\"item\":\"https:\/\/immediatech.net\/personal\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/immediatech.net\/personal\/#website\",\"url\":\"https:\/\/immediatech.net\/personal\/\",\"name\":\"ANTON OSHMIAN\",\"description\":\"Delivery Lead \/ System Architect\",\"publisher\":{\"@id\":\"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/immediatech.net\/personal\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":[\"Person\",\"Organization\"],\"@id\":\"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36\",\"name\":\"antoshby\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/immediatech.net\/personal\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2021\/05\/logoAOb1.png\",\"contentUrl\":\"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2021\/05\/logoAOb1.png\",\"width\":1000,\"height\":1000,\"caption\":\"antoshby\"},\"logo\":{\"@id\":\"https:\/\/immediatech.net\/personal\/#\/schema\/person\/image\/\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph - ANTON OSHMIAN","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/immediatech.net\/personal\/transition-manifesto\/","og_locale":"en_US","og_type":"article","og_title":"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph - ANTON OSHMIAN","og_description":"Move beyond linear LLM chains. Learn how to design resilient, cyclic multi-agent AI architectures and state machines using LangGraph and Python.","og_url":"https:\/\/immediatech.net\/personal\/transition-manifesto\/","og_site_name":"ANTON OSHMIAN","article_published_time":"2026-05-24T13:44:48+00:00","article_modified_time":"2026-06-03T13:59:46+00:00","og_image":[{"width":2240,"height":1260,"url":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg","type":"image\/jpeg"}],"author":"antoshby","twitter_card":"summary_large_image","twitter_misc":{"Written by":"antoshby","Est. reading time":"4 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#article","isPartOf":{"@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/"},"author":{"name":"antoshby","@id":"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36"},"headline":"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph","datePublished":"2026-05-24T13:44:48+00:00","dateModified":"2026-06-03T13:59:46+00:00","mainEntityOfPage":{"@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/"},"wordCount":663,"publisher":{"@id":"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36"},"image":{"@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage"},"thumbnailUrl":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg","articleSection":["insights"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/","url":"https:\/\/immediatech.net\/personal\/transition-manifesto\/","name":"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph - ANTON OSHMIAN","isPartOf":{"@id":"https:\/\/immediatech.net\/personal\/#website"},"primaryImageOfPage":{"@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage"},"image":{"@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage"},"thumbnailUrl":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg","datePublished":"2026-05-24T13:44:48+00:00","dateModified":"2026-06-03T13:59:46+00:00","breadcrumb":{"@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/immediatech.net\/personal\/transition-manifesto\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#primaryimage","url":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg","contentUrl":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2026\/02\/ANTOSH-PRINCEPLE.jpg","width":2240,"height":1260,"caption":"I architect AI automation, web systems & data pipelines that reduce operational costs."},{"@type":"BreadcrumbList","@id":"https:\/\/immediatech.net\/personal\/transition-manifesto\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Personal","item":"https:\/\/immediatech.net\/personal\/"},{"@type":"ListItem","position":2,"name":"From Prompts to State Machines: Designing Cyclic Multi-Agent Workflows with LangGraph"}]},{"@type":"WebSite","@id":"https:\/\/immediatech.net\/personal\/#website","url":"https:\/\/immediatech.net\/personal\/","name":"ANTON OSHMIAN","description":"Delivery Lead \/ System Architect","publisher":{"@id":"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/immediatech.net\/personal\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":["Person","Organization"],"@id":"https:\/\/immediatech.net\/personal\/#\/schema\/person\/7deb50a0f99a4eab9c8f60a08af62c36","name":"antoshby","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/immediatech.net\/personal\/#\/schema\/person\/image\/","url":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2021\/05\/logoAOb1.png","contentUrl":"https:\/\/immediatech.net\/personal\/wp-content\/uploads\/2021\/05\/logoAOb1.png","width":1000,"height":1000,"caption":"antoshby"},"logo":{"@id":"https:\/\/immediatech.net\/personal\/#\/schema\/person\/image\/"}}]}},"_links":{"self":[{"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/posts\/31431","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/comments?post=31431"}],"version-history":[{"count":4,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/posts\/31431\/revisions"}],"predecessor-version":[{"id":31435,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/posts\/31431\/revisions\/31435"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/media\/31300"}],"wp:attachment":[{"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/media?parent=31431"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/categories?post=31431"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/immediatech.net\/personal\/wp-json\/wp\/v2\/tags?post=31431"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}