/CurtisYoung18"}],"camera":{"position":[3.7,0.4314630112634247,0],"target":[0,0,0]},"thumbUrl":"https://assets.intro3d.com/thumbs/VKWg25t3pPT07NKXKkkoCYM7U4gDm3Z6/738c0d10-139e-461e-a190-c1473e22ff6d.webp"},"loading":{"background":"#202020","text":"Loading…","spinnerStyle":"dots","spinnerColor":"#cbf53d","template":"classic"},"portfolio":{"enabled":true,"template":"gallery-horizontal","title":"Works","hint":"Scroll on","bgColor":"#0a0c10","bgOpacity":0.92,"sections":[{"id":"us10o3p","title":"Agent Systems","items":[{"id":"gpo0f5n","name":"Intera","meta":"AgentOps · Customer Service","role":"Product design · Agent delivery","link":"https://curtisyoung18.github.io/products/intera/","tags":["Runtime","Tracing","Evaluation","HITL"],"body":"

A unified operational surface for agent configuration, tool evidence, trace replay, release control, and human handoff.

","coverUrl":"https://assets.intro3d.com/stickers/VKWg25t3pPT07NKXKkkoCYM7U4gDm3Z6/fd2bcb86-c217-464d-be6f-43f05c00d3be.webp"}],"tagline":"Operate what the model does, not just what it says.","awards":["INTERA","2026"],"footer":"Next.js · TypeScript · PostgreSQL","coverUrl":"https://assets.intro3d.com/stickers/VKWg25t3pPT07NKXKkkoCYM7U4gDm3Z6/80d49b86-7837-4a9a-a945-47d4e4a4506c.webp"},{"id":"7g0dal9","title":"Model Research","items":[{"id":"0yx1tsb","name":"AIM","meta":"Model Research · WWW 2025","role":"Research · Productization","link":"https://dl.acm.org/doi/10.1145/3696410.3714737","tags":["LLM","Logits","Adaptation","Evaluation"],"body":"

AIM redistributes logits after training so one model can satisfy different endpoint requirements without retraining.

"}],"tagline":"One model, different end requirements, no retraining.","awards":["ACM WEB","2025"],"footer":"Logits Redistribution · Published Research","coverUrl":"https://assets.intro3d.com/stickers/VKWg25t3pPT07NKXKkkoCYM7U4gDm3Z6/f8a7aae1-e79b-4af0-90e5-1a3692cfd019.webp"},{"id":"tid7dlp","title":"Agent Prototype","items":[{"id":"yzl6y3x","name":"MY_MAP","meta":"Agent Prototype · 2026","role":"AI Solution Design & Delivery","link":"https://curtisyoung18.github.io/products/my-map/","tags":["Next.js","MapLibre","MiniMax","AMap"],"body":"

MY_MAP turns natural-language travel needs into structured routes, stops, and recommendations while keeping tool state visible.

"}],"tagline":"Loose travel intent becomes concrete map actions.","awards":["MY_MAP","2026"],"footer":"Next.js · MapLibre · MiniMax · AMap","coverUrl":"https://assets.intro3d.com/stickers/VKWg25t3pPT07NKXKkkoCYM7U4gDm3Z6/62297151-318f-4fdf-8d76-26dfed9c6a33.webp"},{"id":"xvt1vc3","title":"Delivery Lab","items":[{"id":"ti9mxkx","name":"Questionnaire System","meta":"Public Health · LLM POC","role":"Product design · Workflow automation","link":"https://github.com/CurtisYoung18/v0-poc","tags":["LLM","Questionnaire","Analysis","Government"],"body":"

A system for generating, editing, managing, and analyzing food-poisoning investigation questionnaires with LLM assistance.

"},{"id":"t951ggz","name":"Email Processing","meta":"RAG Pipeline · 2025","role":"Data workflow · Agent integration","link":"https://github.com/CurtisYoung18/Eml_process_platform","tags":["RAG","Email","Knowledge Base","Pipeline"],"body":"

A pipeline that cleans email archives and uploads them into a conversational RAG knowledge base.

"},{"id":"9jac7p3","name":"Service Widget","meta":"AI Customer Service · 2025","role":"Integration · Human handoff","link":"https://v0-i-gaming.vercel.app","tags":["Widget","AI Support","Handoff","Integration"],"body":"

A customer-service widget connecting live chat, AI workflows, and an explicit human-handoff path.

"}],"tagline":"From field workflow to dependable operation.","awards":["PUBLIC HEALTH","RAG","HITL"],"footer":"LLM · Workflow Automation · Integration","coverUrl":"https://assets.intro3d.com/stickers/VKWg25t3pPT07NKXKkkoCYM7U4gDm3Z6/9a21ef35-07a7-43e3-b7e4-d688508c9522.webp"}]},"showBranding":true},"title":"Curtis Young · 3D Resume"}]] 1b:I[91319,[],"IconMark"] 9:{"metadata":[["$","title","0",{"childre