Build production LLM apps using 5+ vector databases and GPT-4o.
Master RAG, prompt engineering, and agentic workflows. Download the ultimate AI developer skill today.
LLM Ops
Vector Search
RAG
Prompt Engineering
AI Agents
About This Skill
Master the modern AI stack with support for 6+ leading models including GPT-4o and Claude. This skill covers 5 vector databases and advanced RAG patterns to build reliable, production-ready AI agents and machine learning pipelines.
Quick Start
1Install the antigravity CLI
2Add the ai-ml-developer skill to your environment
3Configure your LLM API keys and vector database credentials
Example Command
antigravity add ai-ml-developer
Core Capabilities
LLM Integration & Model Management
Connect to GPT-4o, Claude, and open-source models like Llama 3.2 with structured output and function calling.
RAG Systems
Implement advanced retrieval-augmented generation using Pinecone, Qdrant, and semantic chunking strategies.
Multi-Agent Workflows
Design complex agentic systems using LangChain or LlamaIndex for autonomous task execution and orchestration.
Model Evaluation & Optimization
Measure performance with automated metrics and optimize prompts for cost, latency, and reliability.
Usage Examples
Input
Create a RAG pipeline for my PDF documentation.
Output
I will set up a pipeline using LangChain, recursive character chunking, and Pinecone for vector storage.
Before
Vague prompt asking for data extraction without structure.
After
Structured system prompt with explicit JSON schema and few-shot examples for reliable extraction.
Input
Implement a multi-agent research assistant.
Output
Configuring a lead agent to decompose tasks and sub-agents for web search and summarization using LangGraph.
SKILL.md
---
name: ai-ml-developer
description: Use when building LLM applications, RAG systems, or AI agents with production requirements. For implementing vector search, prompt engineering, model evaluation, and ML pipelines. Masters GPT-4o/Claude integration, LangChain orchestration, and statistical methods for business intelligence.
---
# AI/ML Developer
Expert AI/ML engineer specializing in production-grade LLM applications, RAG systems, intelligent agents, and machine learning pipelines. Masters the modern AI stack including vector databases, embedding models, agent frameworks, and evaluation methodologies.
## When to Use This Skill
- Building production LLM applications with GPT-4o, Claude, or Llama
- Implementing RAG systems for document Q&A and knowledge retrieval
- Designing multi-agent workflows with LangChain or LlamaIndex
- Optimizing prompts for reliability, cost, and performance
- Evaluating LLM outputs with automated metrics and human feedback
- Implementing vector search with Pinecone, Qdrant, or pgvector
- Applying ML methods for clustering, classification, or pattern discovery
## Core Capabilities
### 1. LLM Integration & Model Management
**Supported Models:**
- **OpenAI**: GPT-4o, GPT-4o-mini, o1-preview with function calling
- **Anthropic**: Claude 4.5 Sonnet/Haiku with tool use
- **Open Source**: Llama 3.2, Mixtral 8x7B, Qwen 2.5, DeepSeek
- **Local Deployment**: Ollama, vLLM, TGI
### 2. RAG Systems
**Architecture Components:**
- Vector databases: Pinecone, Qdrant, Weaviate, Chroma, pgvector
- Embeddings: text-embedding-3-large, BGE-large, e5-large-v2
- Chunking: Semantic, recursive, document-structure aware
- Retrieval: Dense, sparse (BM25), hybrid search
- Reranking: Cross-encoders, Cohere rerank-3
### 3. Agent Frameworks
**Orchestration Tools:**
- LangChain/LangGraph: Complex workflows, state management
- LlamaIndex: Data-centric AI, advanced retrieval
- CrewAI: Multi-agent collaboration
- OpenAI Assistants: Code interpreter, file search
Frequently Asked Questions
FAQ
Which LLMs and tools are supported by this skill?
It supports major providers like OpenAI (GPT-4o), Anthropic (Claude), and open-source models via Ollama, plus 5+ vector databases like Pinecone, Qdrant, and pgvector.
Who is the target audience for this skill?
This is built for software developers and AI engineers transitioning from basic API calls to production-grade AI system architecture and machine learning pipelines.
How does this differ from basic prompt engineering alternatives?
It focuses on the full engineering lifecycle, including RAG architecture, agent orchestration, and rigorous model evaluation rather than just writing text prompts.
What programming languages and frameworks are supported?
While the concepts are universal, the skill focuses heavily on the Python and TypeScript ecosystems, specifically LangChain, LlamaIndex, and standard ML libraries.
What expected results can I achieve with this skill?
You can expect higher reliability in LLM outputs, significantly reduced hallucination rates through RAG, and scalable multi-agent architectures for complex business logic.
Discussion
Discussion
0 comments
U
On This Page
Trigger Phrases
Use these phrases to activate this skill in your AI coding assistant:
Build a RAG systemOptimize LLM promptsSetup vector databaseCreate AI agent workflowEvaluate model performanceImplement semantic search
