[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-672883-105":59,"doc-detail-672883-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","850-engineering-hours-back-context-aware-agentic-airflow-platform-at-wix","850 Engineering Hours Back - Context-Aware Agentic Airflow Platform at Wix","","A technical talk describing a context-aware agentic Airflow platform at Wix, led by Yarden Wolf. It outlines Wix’s data scale, production Airflow operations, and how Maya—a composite of Wix data engineers—faces key bottlenecks: enforcing internal DAG guardrails, debugging locally when failures occur, and handling critical production alerts. The presentation introduces a two-step approach: establishing a single source of truth via authored agent skills as plugins, and distributing them broadly using an internal agent-agnostic AI Configuration Manager (AICM).",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":30,"@type":76,"position":81},"https://docshare.wps.com/document/technology/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/850-engineering-hours-back-context-aware-agentic-airflow-platform-at-wix/672883/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/850-engineering-hours-back-context-aware-agentic-airflow-platform-at-wix/672883.png","ImageObject",300,407,{"name":92,"@type":93},"Well Done","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-03","2026-11-02",true,{"@type":102,"interactionType":103,"userInteractionCount":14},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What problems does the “Maya” example highlight in maintaining DAGs at Wix?","Question",{"text":112,"@type":113},"Maya’s productivity is limited by three context gaps: ensuring new DAGs meet internal engineering guardrails, figuring out why failures occur locally, and dealing with late-night (2 AM) production alerts.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How does the platform create a single source of truth for agent behavior?",{"text":117,"@type":113},"It authors skills as plugins at the .agents level, so engineers can project those skills into the coding framework they choose while keeping guidance centralized.",{"name":119,"@type":110,"acceptedAnswer":120},"What is AICM and how does it help with agentic coding at scale?",{"text":121,"@type":113},"AICM (AI Configuration Manager) is an internal agent-agnostic framework that versions and distributes agent context (plugins→skills and MCPconfigs) from one source of truth to every repo. It enables publishing once, safe rollback on bad updates, and shared skills across local development and production AirBot.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},672883,1793611882,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":29,"category_name":30,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":14,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":129,"read_time":144},5909892210082,"https://ap-avatar.wpscdn.com/avatar/56001069da27c9f41e8?x-image-process=image/resize,m_fixed,w_180,h_180&k=1789614185077863061","850 Engineering Hours Back, Every Month:  \nContext-Aware Agentic Airflow Platform at Wix  \nYarden Wolf  \nHI, I'M  \nYarden Wolf  \nData & AI Engineer @ Wix  \n● Developer Platform Group (2 Years)  \n● Market Intelligence Group (Current)  \n● Leading AI Initiatives across the Data Engineering Group  \nAGENDA  \nWhere we're headed today  \n01 About Wix & DE’s Group  \nthe company & our data scale  \n03 Production & AirBot  \nresolving incidents autonomously  \n02 Local Development  \ncompliant DAGs & debugging, with context  \n04 Lessons Learned & Next Steps  \nkey takeaways & shift to multi-agent with A2A  \nTHE COMPANY  \nAbout Wix  \nA leading SaaS website builder platform.  \n~280m~10%~45%~30%  \nregistered users globally  \nof all active cms siteson the web run on Wix  \nof market share to sites built with simple website builders  \nBase44  \nAI app builders market share in the U.S.  \nAbout Wix & DE’s Guild  \nLocal Development  \nProduction & AirBot  \nNext Steps  \nWHERE I WORK  \nData @ Wix  \n● 100+ Data Engineers  \n● 50+ Data Scientists  \n● 150+ Data Analysts  \nPeople  \nData  \nAirflow  \n● 7PiB of Data Lake Storage  \n● 20TB of Data processed daily ● 2M SQL queries per day  \n● 2 Production Airflow Clusters ● 8,000+ DAGs in production  \n● 250,000+ Airflow tasks a day  \nAbout Wix & DE’s Guild  \nLocal Development  \nProduction & AirBot  \nNext Steps  \nMEET MAYA  \nMaya vs. The Complexity of Scale  \nMaya M  \na composite of every Wix Data Engineer  \nMaya is good at the job. But three things hold back her productivity.  \n 01 Ensuring new DAGs satisfy all internal engineering guardrails  \n 02 Figuring out why it breaks locally  \n 03 Getting alerted at 2 AM when it breaks in production  \nBATTLE 01  \n“Just write the DAG.”  \nHow hard can it be at our scale? (100+ Data Engineers)  \nMAYA'S PROBLEM  \nThe context gap  \nDeploying a DAG requires alignment with proprietary Wix architecture and internal engineering guardrails:  \nCUSTOM OPERATORS (WIXFLOW)  \nUsing our in-house Airflow operators correctly.  \nDWH MODELING PATTERNS  \nModeling techniques, Trino & Iceberg SQL expressions, partitioning.  \nSTRICT GOVERNANCE RULES  \nDWH Table & column naming conventions and governance rules.  \nAbout Wix & DE’s Guild  \nLocal Development  \nProduction & AirBot  \nNext Steps  \nFIX · STEP 1  \nOne source of truth, for every agent  \nAuthored skills as plugins at the .agents level to project them into whichever framework the engineer chooses.  \nSOURCE OF TRUTH  \n.agents/  \ninstructions · skills  \nAbout Wix & DE’s Guild  \n→  \nCursor  \n.cursor/  \nLocal Development  \nClaude  \n.claude/  \nProduction & AirBot  \nCodex  \n.codex/  \nNext Steps  \nFIX · STEP 2  \nGetting it to all 100+ engineers: AICM  \nAI Configuration Manager—our internal agent agnostic framework for coding-agent configuration. It versions and distributes agent context (plugins-> skills & MCPconfigs) to every repo from a single source of truth.  \n01 One change, distributed to every repo  \nWithout it: updating one skill means a PR in every repo. With AICM: publish once, every repo pulls it.  \n02 Versioned like a package  \nVersions locked per repo. A bad context update rolls back in seconds.  \n03 Engineers aren't the only ones who'll use them  \nThe same team specific skills that guide Maya in her local, become part of the context for AirBot in production.  \nAbout Wix & DE’s Guild  \nLocal Development  \nProduction & AirBot  \nNext Steps","cbCaimsUqhnp6efO","https://ap.wps.com/l/cbCaimsUqhnp6efO","pdf",5169351,36,"English","# About Wix & DE’s Group\n## The company & our data scale\n# Local Development\n## Compliant DAGs & debugging with context\n# Production & AirBot\n## Resolving incidents autonomously\n# Lessons Learned & Next Steps\n## Key takeaways & shift to multi-agent with A2A","[{\"question\":\"What problems does the “Maya” example highlight in maintaining DAGs at Wix?\",\"answer\":\"Maya’s productivity is limited by three context gaps: ensuring new DAGs meet internal engineering guardrails, figuring out why failures occur locally, and dealing with late-night (2 AM) production alerts.\"},{\"question\":\"How does the platform create a single source of truth for agent behavior?\",\"answer\":\"It authors skills as plugins at the .agents level, so engineers can project those skills into the coding framework they choose while keeping guidance centralized.\"},{\"question\":\"What is AICM and how does it help with agentic coding at scale?\",\"answer\":\"AICM (AI Configuration Manager) is an internal agent-agnostic framework that versions and distributes agent context (plugins→skills and MCPconfigs) from one source of truth to every repo. It enables publishing once, safe rollback on bad updates, and shared skills across local development and production AirBot.\"}]","850 Engineering Hours Back - Context-Aware Agentic Airflow Platform at Wix | PDF",91]