The LLM Engineer Roadmap¶
约 97 个字 预计阅读时间不到 1 分钟
converting text into numbers through tokenization, processing these tokens through layers including attention mechanisms, and finally generating new text through various sampling strategies.
1. Running LLMs¶
- LLM APIs
- Open-source LLMs
- Prompt engineering
- Structuring outputs
2. Building a Vector Storage¶
- Ingesting documents
- Splitting documents
- Embedding models
- Vector databases
3. Retrieval Augmented Generation¶
- Orchestrators
- Retrievers
- Memory
- Evaluation
4. Advanced RAG¶
- Query construction
- Agents and tools
- Post-processing
- Program LLMs
5. Agents¶
- Agent fundamentals
- Agent frameworks
- Multi-agents
6. Inference optimization¶
- Flash Attention
- Key-value cache
- Speculative decoding
7. Deploying LLMs¶
- Local deployment
- Demo deployment
- Server deployment
- Edge deployment
8. Securing LLMs¶
- Prompt hacking
- Backdoors
- Defensive measures