Kỹ sư AI
AI Builder
Vietnamese onlyThe learning content is available in Vietnamese only. The surrounding interface follows your language.Xây dựng và triển khai hệ thống AI — fine-tuning, RAG, serving, MLOps.
Module
Kiến trúcCNN, RNN, LSTM và hạn chế; Transformer (self-attention, multi-head, positional encoding) Residual connections, weight init (Xavier, He), batch norm SGD, momentum, Adam, AdamW
Module
LLM & NLPPrompt; GPT vs BERT; tokenization Temperature, top-k/top-p, beam search; KV cache; context window
Module
Fine-tuning & Tối ưuFine-tuning vs prompting; LoRA; QLoRA Quantization, distillation, pruning, mixed precision
Module
RAG & AgentsRAG end-to-end: chunking, embedding, retrieval, re-ranking; FAISS Function calling, structured outputs, orchestration, agentic RAG
Module
Hạ tầng & Vận hànhModel serving, inference optimization, MLOps, containers, monitoring Edge, GPU cost, data pipelines
Module
Đánh giá & Quan sát hệ thống AIOffline/online LLM eval, LLM-as-judge RAG faithfulness / answer-relevance / context-relevance Agent eval, tracing, token cost, canary, prompt-injection defense