17 minutes12-Factor Agents: Patterns of reliable LLM applications — Dex Horthy, HumanLayer
Learn how to design AI workflows that can pause, resume, ask for human judgment and keep business state separate from model guesses.
Advanced techniques — agents, automation, local AI, and workflow orchestration.
42 results
Nothing matches these filtersTry a different category or level, or clear the filters to see everything.
17 minutesLearn how to design AI workflows that can pause, resume, ask for human judgment and keep business state separate from model guesses.
18 minutesSee how approval gates can be implemented as explicit workflow checkpoints rather than informal manual review after something goes wrong.
32 minutesLearn how to introduce AI into localization without removing human ownership of meaning, tone, terminology and final approval.
4 minutesUnderstand why multilingual embeddings matter for private internal search and where local retrieval can reduce data-exposure risk.
66 minutesSee the full beginner workflow for building a small Cursor project, including setup, prompting, debugging, basic Git and the limits of AI-assisted coding.
5 minutesYou can name the three guardrails to set before letting AI write code for you, and know why each one prevents a specific failure.
31 minutesSee how an n8n support workflow drafts retrieval-backed replies and proposes knowledge-base updates, then identify the controls needed before production use.
231 minutesSee how a support agent combines retrieval, escalation and workflow tools before production hardening.
24 minutesDesign a marketing workflow where AI drafts and routes work, but humans keep control over strategy, brand voice and publishing.
30 minutesMap the current marketing-tool landscape before deciding which workflows deserve automation.
26 minutesEvaluate where voice agents might fit sales workflows and where disclosure, consent and escalation become blockers.
30 minutesUnderstand the data-enrichment and personalization pipeline behind AI-assisted outbound without confusing automation with quality.
23 minutesStudy a multi-agent newsletter pipeline and identify where sources, approval and failure handling belong.
24 minutesRecognize the browser-agent action loop, where it helps, and where human confirmation is still required.
2 minutesCompare desktop computer-use behavior with browser-only agents.
11 minutesIdentify the source-selection, question, generated-artefact, and citation-inspection stages in a hosted document assistant.
26 minutesObserve a source-based notebook workflow and identify the retrieval, citation, provenance, and permission checks needed before relying on it.
29 minutesUnderstand generate-score-prune reasoning patterns and when they are too expensive for real work.
25 minutesYou can layer role, structured sections and explicit thinking steps into a prompt without turning chain-of-thought into a ritual.
69 minutesBuild intuition for chunking choices before tuning a real retrieval system.
24 minutesSee how query rewriting, hybrid retrieval, reranking and corrective loops fit into one RAG pipeline.
25 minutesUse the OWASP LLM risk categories to review tool access, output handling and sensitive-data exposure.
11 minutesDistinguish direct and indirect prompt injection and why filtering alone is not enough.
107 minutesLearn the product-builder eval loop: inspect traces, label failures, define criteria, test changes and compare against human judgment.
3 minutesSee the smallest no-code version of a repeatable prompt eval.
26 minutesBuild a first n8n AI-agent workflow while recognizing where tool access, memory and guardrails belong.
92 minutesExtend a simple agent into multi-step workflows with memory, error handling and realistic business integrations.
14 minutesInstall a local model runner, pull a small model and understand the privacy/performance tradeoff before using it for real work.
6 minutesTry local AI through a GUI and compare small-model behavior with hosted frontier models.
20 minutesUnderstand the protocol roles: host, client, server, tools, resources, prompts and transports.
16 minutesExplain what MCP changes in plain language and decide whether a tool connection should use MCP or a simpler integration.
19 minutesCompare flagship, lite, mid-tier and specialized models so routing decisions are based on task fit, cost and latency instead of brand preference.
9 minutesEvaluate model-routing tradeoffs between quality, cost and reliability before adding orchestration complexity.
30 minutesIdentify distinct workflow stages and create testable tool-selection criteria for each one.
14 minutesTurn several use-case ideas into candidate stages that you can test against the article's capability and handoff criteria.
10 minutesChoose an automation platform based on hosting, AI-agent fit, integrations and maintenance tradeoffs instead of product hype.
3 minutesSee the original o1 launch framing and separate that historical account from current implementation guidance.
28 minutesUnderstand one historical account of o1, then compare a direct task specification with other supported prompt variants on your current model and workload.
10 minutesEvaluate C.R.A.F.T. as one candidate structure for generating prompts, using representative cases and explicit acceptance criteria.
20 minutesExtract a candidate prompt structure, then add ownership, versioning, evaluation, limits and fallback fields before review.
8 minutesUse multiple AI future scenarios to plan adoption without betting the company on one forecast.
60 minutesFrame AI adoption around workflow change, capability building and realistic organizational risk.