AI Engineering: Building Production LLM Applications
«AI Engineering: Building Production LLM Applications» è un corso online di RAG e AI engineering di livello avanzato, tenuto da InsightLearn Curated AI su InsightLearn. Il percorso comprende 11 lezioni organizzate in 4 sezioni, per una durata complessiva di 15h 0m. End-to-end guide to shipping LLM applications to production: architecture, latency optimization, caching, safety, and observability.
- Docente: InsightLearn Curated AI
- Categoria: RAG e AI engineering
- Livello: avanzato
- Durata: 15h 0m
- Accesso: gratuito in questa fase di validazione, in cambio di un feedback al termine
- Certificato: certificato di completamento verificabile online
Descrizione del corso
InsightLearn è in fase di sviluppo (work in progress): i contenuti di questa pagina sono pubblicati esclusivamente a fini di verifica e test.
The AI Engineer's complete playbook for production systems. Covers LLM application architecture patterns (gateway, routing, fallbacks), latency optimization techniques (streaming, batching, caching with GPTCache/Redis), prompt injection defense, output parsing and structured generation (Instructor, Outlines), rate limiting and cost management, multi-model routing (LiteLLM), load testing LLM endpoints, blue/green prompt deployments, and building internal developer tooling for AI teams.
Cosa imparerai
- Curated from public resources. End-to-end guide to shipping LLM applications to production: architecture, latency optimization, caching, safety, and observability.
Programma del corso
- LLM gateway pattern: routing, fallbacks, retries (35 min)
- Multi-model routing with LiteLLM (30 min)
- Streaming responses: SSE, WebSocket, and chunked transfer (28 min)
- Prompt caching strategies: semantic and exact match (32 min)
- Batching requests for throughput vs latency trade-offs (25 min)
- Prompt injection: attacks, defenses, and detection (40 min)
- Structured output generation with Instructor and Outlines (35 min)
- Hallucination detection and grounding techniques (38 min)
- Blue/green prompt deployments and canary releases (30 min)
- Load testing LLM endpoints with Locust (28 min)
- Building internal AI developer portals (25 min)
Dalle lezioni
Estratto della trascrizione di «LLM gateway pattern: routing, fallbacks, retries»
Are you building AI applications for production? Here are a few key points about architectural design that you should know. As the AI revolution progresses, more companies are issuing their versions of AI models. These models keep getting updated, fine-tuned, and occasionally receive newer versions. Companies developing AI applications often make calls to these AI models via API. Initially, they used vendor-provided client libraries like Python to access the models. However, with the emergence of AI models, of wrapper libraries like LangChain accessing these models became more sophisticated, allowing for more complex logic to be implemented. This transition made the choice of large language models LLM configurable. Now when companies want to switch from GPT-4O to Claude 3, they can…
Estratto della trascrizione di «Multi-model routing with LiteLLM»
Okay. So recently I've been talking a lot in videos about the whole idea of cheaper models and faster models and being able to use them over some of the more expensive models a lot of the time. And the big reason for doing that is I'm just seeing so many people waste so much money with doing calls to GPT-4 for everything or calls to Claude Opus and stuff like that for lots of things where they don't actually need that kind of model. They would be much better to go for a faster model and a cheaper model. The one challenge with that has always been, well, what if some of your calls to the LLM actually do…
Estratto della trascrizione di «Streaming responses: SSE, WebSocket, and chunked transfer»
There's nothing more frustrating than a poorly timed lag spike. Improving latency is essential for any real-time app, and the two most common approaches are server-sent events and web sockets. But how do you know which one to use? Today, we're breaking down both to help you make the right choice. Server-sent events, or SSE, represent a streamlined approach to server-to-client communication. We're breaking down traditional HTTP requests where the client must constantly poll the server for updates, SSE allows the server to automatically push updates to the client. This is achieved through a single, persistent HTTP connection that remains open as long as needed. The simplicity of this model makes SSE an attractive option for applications that require continuous,…
Altri corsi di RAG e AI engineering
- AI Safety: Alignment, Red-Teaming, Responsible Deployment
- Fine-Tuning LLMs: LoRA, QLoRA, and Full Fine-Tuning
- Function Calling & Tool Use: Multi-Model Production Patterns
- LangSmith: Tracing, Debugging, and Monitoring LLM Apps
- LLM Evals: RAGAS, DeepEval, and Automated Testing
- Local LLMs: Ollama, vLLM, and Self-Hosted AI
Domande frequenti
Quanto dura il corso AI Engineering: Building Production LLM Applications?
Il corso dura 15h 0m in totale, con 11 lezioni suddivise in 4 sezioni, da seguire online al proprio ritmo.
Serve esperienza pregressa?
Il corso è di livello avanzato e non indica requisiti specifici di ingresso.
Il corso rilascia un certificato?
Sì: al completamento del corso ricevi un certificato di completamento verificabile online.