Primitives

fai-basic-prompt-optimization

|

fai-bicep-module-create

Create a reusable Bicep module with typed parameters, conditions, outputs, and AVM alignment.

fai-bicep-module-scaffold

|

fai-boost-prompt

Analyze and refine prompts with specificity scoring, few-shot examples, and constraint encoding — improving accuracy by 20-40% and reducing token spend by 15-25% for RAG and classification tasks.

fai-build-agentic-loops

Build autonomous agent loops with ReAct pattern, tool calling, reflection cycles, and termination conditions — enabling multi-turn reasoning without callback hell or token explosion.

fai-build-bicep-module

Create reusable Bicep modules following AVM patterns with typed parameters, outputs, deployment tests, and registry publishing — standardizing infrastructure across 100+ solution plays.

fai-build-data-lakehouse

No description.

fai-build-docker-image

Build optimized Docker images with multi-stage builds, non-root runtime, health checks, and AI model weight caching — reducing deployment time by 60% and image size by 40%.

fai-build-etl-pipeline

No description.

fai-build-genai-rag

No description.

fai-build-github-workflow

No description.

fai-build-integration-test

No description.

fai-build-kubernetes-manifest

No description.

fai-build-llm-evaluator

Build custom LLM evaluation metrics for groundedness, coherence, relevance, and domain-specific quality — enabling automated CI gates and production quality monitoring for RAG and GenAI systems.

fai-build-nosql-data-model

Build NoSQL data models (Cosmos DB documents, partition keys) for AI workloads with optimal consistency and throughput.

fai-build-prompting-system

Build end-to-end prompting systems using Prompt Flow, versioning, A/B testing, and performance tracking.

fai-build-semantic-search

Implement semantic search with embeddings, reranking, and hybrid search combining BM25 + vector similarity.

fai-build-sql-migration

Migrate SQL schemas to cloud with zero downtime, validation, and rollback automation.

fai-build-terraform-module

Create reusable Terraform modules for multi-cloud AI infrastructure provisioning.

fai-build-test-harness

Build comprehensive test harnesses for LLM applications covering deterministic and non-deterministic flows.

fai-build-tokenizer

Implement custom tokenizers for domain-specific AI models optimizing vocabulary size and encoding efficiency.

fai-build-unit-test

Write effective unit tests for AI functions covering happy path, edge cases, and error handling.

fai-build-vector-store

Provision vector stores (Cosmos DB, Pinecone, Weaviate) with scaling, indexing, and retrieval optimization.

fai-canary-deploy-ai

Implement canary deployment for AI model updates with traffic splitting and rollback on quality drop.