AI Engineering AI Workflows

Workflows in this category

• Retrieval-Augmented Generation — Retrieval-Augmented Generation grounds an LLM in private knowledge by retrieving relevant context from a vector database before generation.
• Agentic Systems — Agentic systems give an LLM the ability to plan, call tools, and execute multi-step tasks autonomously.
• Research Assistant — A research assistant decomposes complex questions, retrieves grounded sources, reasons across them and synthesises cited answers.
• Multimodal Assistant — A multimodal assistant accepts images, audio and text and produces grounded answers, edits or generations.
• Agentic RAG — Agentic RAG upgrades a standard single-shot retrieval pipeline with a planning agent that can reformulate queries, switch retrieval strategies, validate the retrieved evidence against the question and iterate until the answer meets a confidence threshold.
• Graph RAG — Use a knowledge graph alongside vector retrieval to ground LLM answers in entity-rich, relationship-aware context.
• Hybrid Search — Blend lexical (BM25) and semantic retrieval with reranking for state-of-the-art recall and precision.
• Evaluation Pipeline — Continuous offline and online evaluation harness with regression detection, dataset versioning and dashboards.
• MCP Workflow — Expose enterprise systems to LLMs via the Model Context Protocol with auth, scope control and audit logs.
• Multi-Agent Orchestration — Plan, route and supervise multi-agent workflows with typed message contracts and budget circuit-breakers.
• RAG Evaluation — Continuous evaluation harness for RAG pipelines with groundedness, faithfulness and answer-relevance scoring.
• Guardrails Pipeline — Composable pre/post-generation guardrail layer covering PII, toxicity, jailbreaks and policy violations.
• Model Routing — Dynamic model routing across providers and tiers based on intent, cost ceiling and latency SLA.
• Synthetic Data Generation — Generate, evaluate and govern synthetic datasets for evals, fine-tuning and offline testing.
• Fine-Tuning Pipeline — End-to-end fine-tuning pipeline with data prep, eval, training and deploy stages plus rollback gates.
• Prompt Management — Versioned prompt registry with diff review, eval-gated promotion and per-environment rollout.
• Vector Index Management — Operational playbook for vector-index lifecycle: schema, chunking, reindex, freshness and tiered storage.
• Semantic Cache — Production semantic caching layer with similarity gating, TTL policy and cache-quality monitoring.
• LLM Observability — End-to-end observability stack for LLM apps: traces, evals, cost, drift and incident response.
• Benchmark Bot — Benchmark Bot is a conversational RAG application over the ModelRefs canonical benchmark graph.

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to AI Engineering AI Workflows.

• Browse models
• Review providers
• Inspect benchmarks
• Explore workflows
• Learn AI concepts
• Read decision guides
• Review trust and methodology
• Start a decision path