[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84664-en":3,"doc-seo-84664-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84664,4810365810221,"Aurora","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","LLMoxie Exploring Agentic AI for Scientific Software Development","LLMoxie is an institutional agentic AI platform designed for scientific software development, emphasizing reproducibility, auditability, and safe handling of sensitive research data. It provides a three-tier architecture enabling multi-cloud and on-premise inference, with a LiteLLM/MLflow control plane for authentication, budgeting, PII masking, and observability. An application layer supports AI coding agents via an open-source RSE-Plugins ecosystem that encodes RSE expertise as Plugin-Agent-Skill hierarchies and supports a six-phase research workflow. Over 20 months in a university RSE center, the work addresses infrastructure, governance, and process challenges and converts generic code generation into domain-aware collaboration with provenance of technical reasoning.","LLMoxie: Exploring Agentic AI for Scientific Software  \nDevelopment  \nLandung Setiawan  \neScience Institute University of Washington Seattle, Washington, USA[landungs@uw.edu](landungs@uw.edu)  \nAnant Mittal  \neScience Institute University of Washington Seattle, Washington, USA [anmittal@uw.edu](anmittal@uw.edu)  \nCordero Core  \neScience Institute University of Washington Seattle, Washington, USA [cdcore@uw.edu](cdcore@uw.edu)  \narXiv :2607 .02703v 1 [ cs . SE] 2 Jul 2026  \nAnshul Tambay  \neScience Institute University of Washington Seattle, Washington, USA [anshul37@uw.edu](anshul37@uw.edu)  \nCarlos Garcia Jurado Suarez  \neScience Institute University of Washington Seattle, Washington, USA [carlosg@uw.edu](carlosg@uw.edu)  \nDavid A. C. Beck  \neScience Institute Dept of Chemical Engineering University of Washington Seattle, Washington, USA[dacb@uw.edu](dacb@uw.edu)  \nAndrew J. Connolly  \neScience Institute Dept of Astronomy University of Washington Seattle, Washington, USA [ajc@astro.washington.edu](ajc@astro.washington.edu)  \nVani Mandava  \neScience Institute University of Washington Seattle, Washington, USA [vani1@uw.edu](vani1@uw.edu)  \nFigure 1: An RSE-oriented scientific software development lifecycle. Conventional engineering stages (Spec, Dev, Test, Ship) are bracketed by science-specific phases: upstream framing of the research problem, software assumptions, and goals clarification, and downstream handoff and community building. Scientists drive the bookend phases while research software engineers carry the work across the middle, illustrating where AI-assisted coding agents must integrate to support scientific practice.  \nAbstract  \nIn this paper, we describe LLMoxie, an institutional AI platform whose three-tiered architecture supports multi-cloud and on-premise inference, a LiteLLM/MLflow control plane for authentication, budgeting, PII masking, and observability, and  \nan application augmentation layer for AI coding agents. Layered ontop, an open-source RSE-Plugins ecosystem encodes accumulated RSE knowledge as a Plugin-Agent-Skill hierarchy spanning scientific Python practice, domain-specific knowledge, a sixphase research-and-implement workflow, and project lifecycle  \nmanagement. Scientific software is judged less by raw code quality than by whether it can be cited, audited, reproduced, and extended. Off-the-shelf AI coding agents, optimized against commercial software benchmarks, are poorly calibrated for this setting: they ignore the conventions of the scientific Python libraries they invoke, mishandle sensitive or embargoed data, and leave decision trails that are difficult to reconstruct after the fact. We report on twenty months of practice at a university-based research software engineering (RSE) center, where RSEs embedded across astronomy, earth and climate science, agriculture, and health projects worked to close this gap. We characterize the recurring infrastructure, governance, and process challenges of adopting Agentic AI inside a multi-domain RSE center, describe the platform and plugin design, and distill operational lessons from real scientific software deployments. Together, the platform and plugins shift AI coding agents from generic code generators into domain-aware collaborators that respect community norms and produce auditable provenance of technical reasoning.  \n1 Introduction  \nOur research software engineering center, Scientific Software Engineering Center, is a professional software engineering organization embedded within the University of Washington, Seattle. It serves as a campus-wide collaborator for domain scientists, translating research questions into durable software systems by clarifying scientific goals, designing reproducible workflows, testing underlying assumptions, and delivering maintainable tools. Since its establishment, the center has delivered more than 20 multi-institution scientific software projects, convened workshops and events reaching over 1,000 participants, sus","cbCaii5204n6Lclx","https://ap.wps.com/l/cbCaii5204n6Lclx","pdf",4340997,1,9,"English","en",105,"# Introduction\n## Scientific software as a distinctive setting for AI-assisted development\n## LLMoxie platform architecture and control plane\n## Application augmentation with RSE-Plugins ecosystem\n## Operational challenges and lessons from RSE practice","[{\"question\":\"What is LLMoxie and what problem does it target in scientific software development?\",\"answer\":\"LLMoxie is an institutional AI platform that supports agentic AI coding while respecting scientific software requirements such as citation, auditing, reproducibility, and extension. It targets the gap where commercial AI coding agents are poorly calibrated for research-code conventions and governance needs.\"},{\"question\":\"How does LLMoxie handle multi-cloud or on-premise inference and operational controls?\",\"answer\":\"LLMoxie uses a three-tiered architecture for multi-cloud and on-premise inference and a LiteLLM/MLflow control plane. The control plane supports authentication, budgeting, PII masking, and observability.\"},{\"question\":\"How do the RSE-Plugins and the Plugin-Agent-Skill hierarchy help improve AI coding agents?\",\"answer\":\"The open-source RSE-Plugins ecosystem encodes accumulated RSE knowledge into a Plugin-Agent-Skill hierarchy. It spans scientific Python practice, domain knowledge, and a six-phase research-and-implement workflow to make agents domain-aware and produce auditable provenance.\"}]",1784197548,23,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"llmoxie-exploring-agentic-ai-for-scientific-software-development","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/llmoxie-exploring-agentic-ai-for-scientific-software-development/84664/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is LLMoxie and what problem does it target in scientific software development?","Question",{"text":75,"@type":76},"LLMoxie is an institutional AI platform that supports agentic AI coding while respecting scientific software requirements such as citation, auditing, reproducibility, and extension. It targets the gap where commercial AI coding agents are poorly calibrated for research-code conventions and governance needs.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does LLMoxie handle multi-cloud or on-premise inference and operational controls?",{"text":80,"@type":76},"LLMoxie uses a three-tiered architecture for multi-cloud and on-premise inference and a LiteLLM/MLflow control plane. The control plane supports authentication, budgeting, PII masking, and observability.",{"name":82,"@type":73,"acceptedAnswer":83},"How do the RSE-Plugins and the Plugin-Agent-Skill hierarchy help improve AI coding agents?",{"text":84,"@type":76},"The open-source RSE-Plugins ecosystem encodes accumulated RSE knowledge into a Plugin-Agent-Skill hierarchy. It spans scientific Python practice, domain knowledge, and a six-phase research-and-implement workflow to make agents domain-aware and produce auditable provenance.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]