[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79S58152Q5BWW0CESTKBCYM":3,"skill-same:01M79S58152Q5BWW0CESTKBCYM":51},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":30,"commitSha":31,"skillDirectory":32,"forksCount":14,"repositoryHtmlUrl":33,"repositoryHomepageUrl":6,"ownerHtmlUrl":34,"ownerBlogUrl":6,"ownerDescription":35,"updatedAt":36,"install":37,"resources":50},"01M79S58152Q5BWW0CESTKBCYM","ai-ml-architect",null,"Entry point for AI\u002FML system architecture — choosing between prompting, retrieval, fine-tuning and training; deciding where the model boundary sits; costing and gating inference; and the evaluation discipline that must exist before any of it ships. Use when planning an LLM product, a RAG system, an agent, a model serving stack, or an ML platform.","SriAkhilSJ\u002FPulseAIRepo","SriAkhilSJ","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F206201543?v=4","01M79RWFD1Q09B89SXTTKKJY6Z","Sri Akhil SJ","01M79RWFE26NPG0BZ1FRNX376H",0,"MIT","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fai-ml-architect\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ai-ml-architect\ndescription: Entry point for AI\u002FML system architecture — choosing between prompting, retrieval, fine-tuning and training; deciding where the model boundary sits; costing and gating inference; and the evaluation discipline that must exist before any of it ships. Use when planning an LLM product, a RAG system, an agent, a model serving stack, or an ML platform.\nlicense: MIT\ncompatibility: Any provider or self-hosted stack\nmetadata:\n  crusor-domain: ai-ml-architect\n  crusor-path: ai-ml-architect\n  crusor-origin: original\n  crusor-sources: \"platform.openai.com\u002Fdocs; docs.anthropic.com; huggingface.co\u002Fdocs; arxiv.org\u002Fabs\u002F2005.11401\"\n  crusor-triggers: \"llm, rag, fine tune, embedding, prompt, agent, inference, model serving, mlops, eval, vector database, training\"\n  crusor-verified-on: \"2026-10-30\"\n---\n\n# AI\u002FML Architect\n\nThe model is **a dependency with non-determinism, latency variance, a price per\ntoken, and a quality contract you must measure**. Everything in this domain\nfollows from treating it that way.\n\n## The shelf\n\nEight shelves, 30 books.\n\n| Shelf | Books |\n|---|---|\n| `providers-architect\u002F` | `openai` · `anthropic` · `gemini` · `open-weights` |\n| `llm-architect\u002F` | `llm-app` · `prompting` · `structured-output` · `model-selection` · `token-economics` |\n| `retrieval-architect\u002F` | `rag` · `chunking` · `embeddings` · `vector-store` · `hybrid-search` |\n| `evaluation-architect\u002F` | `evals` · `judge-calibration` · `red-teaming` |\n| `training-architect\u002F` | `fine-tuning` · `classic-ml` · `recommender` · `vision` |\n| `serving-architect\u002F` | `inference-serving` · `gpu-capacity` · `quantization` · `speech` |\n| `mlops-architect\u002F` | `mlops-platform` · `feature-store` · `model-monitoring` |\n| `agent-architect\u002F` | `agent` · `multi-agent` |\n\nRead the crown, then the one or two books the plan actually needs. Do not read\nthe whole shelf to avoid a decision.\n\n## When to use\n\n- Any plan that puts a model in the request path of a product.\n- Choosing between API, self-host, fine-tune, or classic ML.\n- An existing AI feature that is expensive, slow, or unmeasurable.\n\n## When NOT to use\n\n- Deterministic logic solves it: rules, SQL, a regex, a lookup table.\n- The task is \"add a chatbot\" with no defined job to be done.\n- Nobody can state what a correct answer looks like — that is an eval problem\n  first (`evals-architect`), not an architecture problem.\n\n## Decision 1 — Is a model the right tool\n\n| Signal | Better answer |\n|---|---|\n| exact, auditable rule | deterministic code |\n| classification over known labels | classic ML or a small model |\n| free-form text transformation | LLM |\n| answer must cite a fixed corpus | retrieval, not memory |\n| high-volume, narrow format | fine-tune or smaller model |\n| truly open-ended reasoning | large model, with evals |\n\nNon-determinism is a product decision: anything user-visible must have a stated\ntolerance for variance and a human path when confidence is low.\n\n## Decision 2 — Where the boundary sits\n\n- **Prompt-only** — fastest to ship, no training, weakest on format stability.\n- **Retrieval** — grounds answers in your data; adds a retrieval system to own.\n- **Fine-tune** — buys format and tone, not knowledge; needs a dataset and an\n  eval set, plus retraining as a lifecycle.\n- **Train from scratch** — rarely justified outside a domain with proprietary\n  data and volume.\n- Escalate in that order. Each step adds a system you must operate.\n\n## Decision 3 — The envelope\n\nEvery design doc states, numerically:\n\n- quality: eval score on the golden set, and the floor below which it is a\n  regression;\n- latency: p50\u002Fp95 end to end, including every model call in a chain;\n- cost: per request, per user, per month at projected volume;\n- failure: what the user sees when the provider is down or rate-limits.\n\nA chain of five model calls multiplies all four. Count the hops.\n\n## Decision 4 — Data and privacy\n\n- What leaves your boundary, to whom, under which retention terms.\n- PII in prompts is a data-flow decision, not a prompt-engineering one.\n- Tenant isolation: do not mix documents across tenants in one index.\n- Logging prompts is logging user content — apply the same retention rules.\n\n## Decision 5 — Build order\n\n1. Define the job and the golden set.\n2. Baseline: best prompt on the largest model. Measure.\n3. Add retrieval or a smaller model only where the baseline fails.\n4. Add evals to CI before adding features.\n5. Optimise cost and latency against a measured baseline, never by guessing.\n\n## Failure modes at scale\n\n- Shipping without an eval set: quality debates settled by vibes.\n- Fine-tuning to teach facts that retrieval should supply.\n- Prompt chains whose cost and p95 were never summed.\n- One provider, no fallback, no timeout, no budget cap.\n- \"It works\" demonstrated with three cherry-picked examples.\n\n## Verification receipts\n\n1. Job statement plus a golden set with expected outputs.\n2. Baseline eval score, and the regression floor.\n3. Cost and latency model per request at projected volume.\n4. Data-flow diagram: what leaves the boundary and under what terms.\n5. Fallback behaviour verified by simulating a provider outage.\n","# AI\u002FML Architect\n\nThe model is **a dependency with non-determinism, latency variance, a price per\ntoken, and a quality contract you must measure**. Everything in this domain\nfollows from treating it that way.\n\n## The shelf\n\nEight shelves, 30 books.\n\n| Shelf | Books |\n|---|---|\n| `providers-architect\u002F` | `openai` · `anthropic` · `gemini` · `open-weights` |\n| `llm-architect\u002F` | `llm-app` · `prompting` · `structured-output` · `model-selection` · `token-economics` |\n| `retrieval-architect\u002F` | `rag` · `chunking` · `embeddings` · `vector-store` · `hybrid-search` |\n| `evaluation-architect\u002F` | `evals` · `judge-calibration` · `red-teaming` |\n| `training-architect\u002F` | `fine-tuning` · `classic-ml` · `recommender` · `vision` |\n| `serving-architect\u002F` | `inference-serving` · `gpu-capacity` · `quantization` · `speech` |\n| `mlops-architect\u002F` | `mlops-platform` · `feature-store` · `model-monitoring` |\n| `agent-architect\u002F` | `agent` · `multi-agent` |\n\nRead the crown, then the one or two books the plan actually needs. Do not read\nthe whole shelf to avoid a decision.\n\n## When to use\n\n- Any plan that puts a model in the request path of a product.\n- Choosing between API, self-host, fine-tune, or classic ML.\n- An existing AI feature that is expensive, slow, or unmeasurable.\n\n## When NOT to use\n\n- Deterministic logic solves it: rules, SQL, a regex, a lookup table.\n- The task is \"add a chatbot\" with no defined job to be done.\n- Nobody can state what a correct answer looks like — that is an eval problem\n  first (`evals-architect`), not an architecture problem.\n\n## Decision 1 — Is a model the right tool\n\n| Signal | Better answer |\n|---|---|\n| exact, auditable rule | deterministic code |\n| classification over known labels | classic ML or a small model |\n| free-form text transformation | LLM |\n| answer must cite a fixed corpus | retrieval, not memory |\n| high-volume, narrow format | fine-tune or smaller model |\n| truly open-ended reasoning | large model, with evals |\n\nNon-determinism is a product decision: anything user-visible must have a stated\ntolerance for variance and a human path when confidence is low.\n\n## Decision 2 — Where the boundary sits\n\n- **Prompt-only** — fastest to ship, no training, weakest on format stability.\n- **Retrieval** — grounds answers in your data; adds a retrieval system to own.\n- **Fine-tune** — buys format and tone, not knowledge; needs a dataset and an\n  eval set, plus retraining as a lifecycle.\n- **Train from scratch** — rarely justified outside a domain with proprietary\n  data and volume.\n- Escalate in that order. Each step adds a system you must operate.\n\n## Decision 3 — The envelope\n\nEvery design doc states, numerically:\n\n- quality: eval score on the golden set, and the floor below which it is a\n  regression;\n- latency: p50\u002Fp95 end to end, including every model call in a chain;\n- cost: per request, per user, per month at projected volume;\n- failure: what the user sees when the provider is down or rate-limits.\n\nA chain of five model calls multiplies all four. Count the hops.\n\n## Decision 4 — Data and privacy\n\n- What leaves your boundary, to whom, under which retention terms.\n- PII in prompts is a data-flow decision, not a prompt-engineering one.\n- Tenant isolation: do not mix documents across tenants in one index.\n- Logging prompts is logging user content — apply the same retention rules.\n\n## Decision 5 — Build order\n\n1. Define the job and the golden set.\n2. Baseline: best prompt on the largest model. Measure.\n3. Add retrieval or a smaller model only where the baseline fails.\n4. Add evals to CI before adding features.\n5. Optimise cost and latency against a measured baseline, never by guessing.\n\n## Failure modes at scale\n\n- Shipping without an eval set: quality debates settled by vibes.\n- Fine-tuning to teach facts that retrieval should supply.\n- Prompt chains whose cost and p95 were never summed.\n- One provider, no fallback, no timeout, no budget cap.\n- \"It works\" demonstrated with three cherry-picked examples.\n\n## Verification receipts\n\n1. Job statement plus a golden set with expected outputs.\n2. Baseline eval score, and the regression floor.\n3. Cost and latency model per request at projected volume.\n4. Data-flow diagram: what leaves the boundary and under what terms.\n5. Fallback behaviour verified by simulating a provider outage.\n",{"name":5,"license":15,"metadata":24,"description":7,"compatibility":29},{"crusor-path":5,"crusor-domain":5,"crusor-origin":25,"crusor-sources":26,"crusor-triggers":27,"crusor-verified-on":28},"original","platform.openai.com\u002Fdocs; docs.anthropic.com; huggingface.co\u002Fdocs; arxiv.org\u002Fabs\u002F2005.11401","llm, rag, fine tune, embedding, prompt, agent, inference, model serving, mlops, eval, vector database, training","2026-10-30","Any provider or self-hosted stack","main","8fa4e240be65b98472c9b2d3a505ade2aea1a5c6","skills\u002Fai-ml-architect","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ","Studying in CMR institute of Technology\r\n2024 - 2028","2026-11-13 21:02:23",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":49},[39,42,45],{"tool":40,"command":41},"npx","npx skills add SriAkhilSJ\u002FPulseAIRepo --skill ai-ml-architect",{"tool":43,"command":44},"bunx","bunx skills add SriAkhilSJ\u002FPulseAIRepo --skill ai-ml-architect",{"tool":46,"command":47},"pnpm","pnpm dlx skills add SriAkhilSJ\u002FPulseAIRepo --skill ai-ml-architect","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fai-ml-architect\u002FSKILL.md",true,[],[52,59,66,73,80,87],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M79TTAJNACCNXCEN3649CMMD","interaction-patterns-architect","Architect-level interaction design in code — choosing established patterns over invented ones, feedback and latency behaviour, destructive-action safeguards, navigation and overlay conventions, and the state model behind every interactive surface. Use when a screen has to behave, not just look.","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fwebsite-architect\u002Fui-ux-implementation-architect\u002Finteraction-patterns-architect\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M79TT2ESKB1TKZBYY9KYJHV9","design-system-ops-architect","Architect-level operation of a design system in code — token pipelines and versioning, component API and variant policy, the catalogue as the contract, adoption metrics, and the governance that stops a system from forking into three. Use when a component library must survive several teams.","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fwebsite-architect\u002Fui-ux-implementation-architect\u002Fdesign-system-ops-architect\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M79TSV0PWGNK3PRX9P4JDR9N","design-handoff-architect","Architect-level design handoff — what must be specified before implementation starts, tokens over raw values, responsive and content-limit contracts, state matrices, and the review loop that catches drift between design and code. Use when a design is about to become code.","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fwebsite-architect\u002Fui-ux-implementation-architect\u002Fdesign-handoff-architect\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M79TSJS9BFXNP696FMJ9GKDA","content-microcopy-architect","Architect-level UI content — where strings live, voice and terminology consistency, error and empty-state copy as part of the interface, translation-ready authoring, and the content model a product needs before it multiplies screens. Use when the words in the interface matter as much as the pixels.","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fwebsite-architect\u002Fui-ux-implementation-architect\u002Fcontent-microcopy-architect\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M79TSCNBST3N1M68V1KPH9HQ","ui-ux-implementation-architect","Architect-level direction for turning a design into working UI — the handoff contract, component ownership versus screen-level layout, interaction and state conventions, design QA gates, and the review habits that keep an implementation faithful without frozen pixels. Use before any specific UI implementation skill.","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fwebsite-architect\u002Fui-ux-implementation-architect\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M79TS4N6A4D2QQ433VXS5D0R","test-strategy-architect","Architect-level test strategy — allocating assertions across unit, integration, contract and end-to-end layers, designing for testability through dependency boundaries, fixture and data strategy, and the flake and coverage policies that keep a suite trustworthy. Use when defining how a codebase is tested.","https:\u002F\u002Fgithub.com\u002FSriAkhilSJ\u002FPulseAIRepo\u002Fblob\u002Fmain\u002Fskills\u002Fwebsite-architect\u002Ftesting-and-deployment-architect\u002Ftest-strategy-architect\u002FSKILL.md",[],[]]