🚀 Agentic AI Engineer — 3-Month Crash Course

curl -fsSL https://ollama.ai/install.sh | sh ollama pull llama3.2 # 3B — fast, general purpose ollama pull qwen2.5:7b # 7B — excellent for coding & agents ollama pull mistral # 7B — great for tool use ollama pull nomic-embed-text # embeddings — replaces OpenAI text-embedding-3 ollama pull llava # multimodal vision # Runs on http://localhost:11434 (OpenAI-compatible API)

# Qdrant — vector DB (replaces Pinecone) docker run -d -p 6333:6333 qdrant/qdrant # UI → http://localhost:6333/dashboard # Langfuse — LLM observability (replaces LangSmith) git clone https://github.com/langfuse/langfuse && cd langfuse docker-compose up -d # UI → http://localhost:3000 # Weaviate — hybrid search vector DB docker run -d -p 8080:8080 semitechnologies/weaviate:latest # Arize Phoenix — local LLM eval (no account needed) pip install arize-phoenix python -m phoenix.server.main # UI → http://localhost:6006 # MLflow — experiment tracking (replaces W&B) pip install mlflow && mlflow ui # UI → http://localhost:5000

pip install langchain langchain-community langchain-ollama langchain-google-genai pip install langgraph langchain-core pip install llama-index llama-index-llms-ollama llama-index-embeddings-ollama pip install chromadb qdrant-client weaviate-client faiss-cpu pip install ragas trulens-eval arize-phoenix pip install sentence-transformers pip install fastapi uvicorn langserve sse-starlette pip install unsloth axolotl peft transformers trl accelerate pip install docling marker-pdf paddleocr pip install pymupdf pdfplumber instructor pip install mlflow litellm pip install crewai google-adk autogen-agentchat pip install guardrails-ai nemo-guardrails

from langchain_ollama import ChatOllama, OllamaEmbeddings # Test local LLM llm = ChatOllama(model="llama3.2") print(llm.invoke("Say hello in one sentence!").content) # Test local embeddings (replaces OpenAI) emb = OllamaEmbeddings(model="nomic-embed-text") v = emb.embed_query("Hello world") print(f"Embedding dim: {len(v)}") # 768 # Test Gemini free API import os; os.environ["GOOGLE_API_KEY"] = "your-key-from-aistudio" from langchain_google_genai import ChatGoogleGenerativeAI gem = ChatGoogleGenerativeAI(model="gemini-1.5-flash") print(gem.invoke("Say hello!").content) # Test Qdrant (after docker run) from qdrant_client import QdrantClient q = QdrantClient(url="http://localhost:6333") print(q.get_collections()) # {"result": {"collections": []}}