embedding-strategies
Guide to selecting and optimizing embedding models for vector search applications.
About this skill
This skill provides a comprehensive guide for AI agents to effectively select and optimize embedding models crucial for various vector search applications, especially in the context of Retrieval-Augmented Generation (RAG) systems. It covers best practices for choosing the right embedding model, strategies for optimizing text chunking, and insights into fine-tuning embedding models to enhance performance. The skill enables an AI agent to clarify user goals, apply relevant expertise, and provide actionable, verified steps for implementing robust embedding solutions.
Best use case
Utilize this skill to receive expert guidance on choosing appropriate embedding models for RAG applications, developing efficient text chunking strategies to improve vector search relevance, and understanding the principles of fine-tuning embedding models for specific use cases.
Guide to selecting and optimizing embedding models for vector search applications.
A detailed strategic plan, a set of recommendations for selecting and optimizing embedding models, or actionable steps for implementing specific embedding strategies (e.g., chunking, fine-tuning) tailored to the user's vector search application goals.
Practical example
Example input
I'm building a RAG system for medical documents. Can you help me select the best embedding model and chunking strategy to maximize retrieval accuracy?
Example output
To optimize your RAG system for medical documents, we'll focus on model selection, chunking, and potential fine-tuning. **1. Embedding Model Selection:** Given the specialized nature of medical documents, a general-purpose model might not be sufficient. Consider: - **Domain-specific models:** Look for models pre-trained or fine-tuned on biomedical text (e.g., BioBERT, ClinicalBERT, or models from Hugging Face's medical NLP collections). - **Evaluation:** Test several candidates using metrics relevant to your downstream task (e.g., retrieval precision, recall on a medical Q&A dataset). Benchmarking with a small, representative dataset is crucial. - **Model size vs. performance:** Balance the need for accuracy with computational constraints. **2. Chunking Strategy Optimization:** Effective chunking is vital for retrieval quality. For medical documents: - **Semantic Chunking:** Aim to keep semantically related sentences or paragraphs together. Avoid splitting critical concepts. - **Fixed-size with overlap:** A common starting point is chunks of 200-500 tokens with 10-20% overlap. Experiment with these values. - **Section-aware chunking:** If documents have clear headings (e.g., 'Patient History', 'Diagnosis', 'Treatment Plan'), respect these boundaries. Each section could be a chunk, or further sub-chunked. - **Metadata inclusion:** Embed relevant metadata (e.g., document title, section heading, date) alongside the text chunk to provide context during retrieval. **3. Fine-tuning (if necessary):** If off-the-shelf models or initial strategies underperform: - **Dataset:** Create a small, high-quality dataset of (query, relevant_document_chunk) pairs from your medical domain. - **Method:** Use techniques like contrastive learning (e.g., training with positive and negative pairs) to fine-tune a chosen base embedding model, teaching it to better distinguish relevant medical concepts. **Actionable Steps:** 1. Identify 2-3 domain-specific or robust general-purpose embedding models. 2. Define a preliminary chunking strategy (e.g., 400 tokens with 50-token overlap). 3. Set up a small evaluation dataset of typical medical queries and expected document chunks. 4. Implement and benchmark the models and chunking strategies. 5. Iterate on chunking parameters or consider fine-tuning if initial results are not satisfactory. Would you like to explore specific model recommendations or a deeper dive into chunking techniques?
When to use this skill
- Choosing embedding models for RAG
- Optimizing chunking strategies
- Fine-tuning embedding models for improved performance
When not to use this skill
- The task is unrelated to embedding strategies
- You need a different domain or tool outside this scope
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/embedding-strategies/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How embedding-strategies Compares
| Feature / Agent | embedding-strategies | Standard Approach |
|---|---|---|
| Platform Support | Claude | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | easy | N/A |
Frequently Asked Questions
What does this skill do?
Guide to selecting and optimizing embedding models for vector search applications.
Which AI agents support this skill?
This skill is designed for Claude.
How difficult is it to install?
The installation complexity is rated as easy. You can find the installation instructions above.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
Related Guides
Best AI Skills for Claude
Explore the best AI skills for Claude and Claude Code across coding, research, workflow automation, documentation, and agent operations.
AI Agents for Coding
Browse AI agent skills for coding, debugging, testing, refactoring, code review, and developer workflows across Claude, Cursor, and Codex.
ChatGPT vs Claude for Agent Skills
Compare ChatGPT and Claude for AI agent skills across coding, writing, research, and reusable workflow execution.
SKILL.md Source
# Embedding Strategies
Guide to selecting and optimizing embedding models for vector search applications.
## Do not use this skill when
- The task is unrelated to embedding strategies
- You need a different domain or tool outside this scope
## Instructions
- Clarify goals, constraints, and required inputs.
- Apply relevant best practices and validate outcomes.
- Provide actionable steps and verification.
- If detailed examples are required, open `resources/implementation-playbook.md`.
## Use this skill when
- Choosing embedding models for RAG
- Optimizing chunking strategies
- Fine-tuning embeddings for domains
- Comparing embedding model performance
- Reducing embedding dimensions
- Handling multilingual content
## Core Concepts
### 1. Embedding Model Comparison
| Model | Dimensions | Max Tokens | Best For |
|-------|------------|------------|----------|
| **text-embedding-3-large** | 3072 | 8191 | High accuracy |
| **text-embedding-3-small** | 1536 | 8191 | Cost-effective |
| **voyage-2** | 1024 | 4000 | Code, legal |
| **bge-large-en-v1.5** | 1024 | 512 | Open source |
| **all-MiniLM-L6-v2** | 384 | 256 | Fast, lightweight |
| **multilingual-e5-large** | 1024 | 512 | Multi-language |
### 2. Embedding Pipeline
```
Document → Chunking → Preprocessing → Embedding Model → Vector
↓
[Overlap, Size] [Clean, Normalize] [API/Local]
```
## Templates
### Template 1: OpenAI Embeddings
```python
from openai import OpenAI
from typing import List
import numpy as np
client = OpenAI()
def get_embeddings(
texts: List[str],
model: str = "text-embedding-3-small",
dimensions: int = None
) -> List[List[float]]:
"""Get embeddings from OpenAI."""
# Handle batching for large lists
batch_size = 100
all_embeddings = []
for i in range(0, len(texts), batch_size):
batch = texts[i:i + batch_size]
kwargs = {"input": batch, "model": model}
if dimensions:
kwargs["dimensions"] = dimensions
response = client.embeddings.create(**kwargs)
embeddings = [item.embedding for item in response.data]
all_embeddings.extend(embeddings)
return all_embeddings
def get_embedding(text: str, **kwargs) -> List[float]:
"""Get single embedding."""
return get_embeddings([text], **kwargs)[0]
# Dimension reduction with OpenAI
def get_reduced_embedding(text: str, dimensions: int = 512) -> List[float]:
"""Get embedding with reduced dimensions (Matryoshka)."""
return get_embedding(
text,
model="text-embedding-3-small",
dimensions=dimensions
)
```
### Template 2: Local Embeddings with Sentence Transformers
```python
from sentence_transformers import SentenceTransformer
from typing import List, Optional
import numpy as np
class LocalEmbedder:
"""Local embedding with sentence-transformers."""
def __init__(
self,
model_name: str = "BAAI/bge-large-en-v1.5",
device: str = "cuda"
):
self.model = SentenceTransformer(model_name, device=device)
def embed(
self,
texts: List[str],
normalize: bool = True,
show_progress: bool = False
) -> np.ndarray:
"""Embed texts with optional normalization."""
embeddings = self.model.encode(
texts,
normalize_embeddings=normalize,
show_progress_bar=show_progress,
convert_to_numpy=True
)
return embeddings
def embed_query(self, query: str) -> np.ndarray:
"""Embed a query with BGE-style prefix."""
# BGE models benefit from query prefix
if "bge" in self.model.get_sentence_embedding_dimension():
query = f"Represent this sentence for searching relevant passages: {query}"
return self.embed([query])[0]
def embed_documents(self, documents: List[str]) -> np.ndarray:
"""Embed documents for indexing."""
return self.embed(documents)
# E5 model with instructions
class E5Embedder:
def __init__(self, model_name: str = "intfloat/multilingual-e5-large"):
self.model = SentenceTransformer(model_name)
def embed_query(self, query: str) -> np.ndarray:
return self.model.encode(f"query: {query}")
def embed_document(self, document: str) -> np.ndarray:
return self.model.encode(f"passage: {document}")
```
### Template 3: Chunking Strategies
```python
from typing import List, Tuple
import re
def chunk_by_tokens(
text: str,
chunk_size: int = 512,
chunk_overlap: int = 50,
tokenizer=None
) -> List[str]:
"""Chunk text by token count."""
import tiktoken
tokenizer = tokenizer or tiktoken.get_encoding("cl100k_base")
tokens = tokenizer.encode(text)
chunks = []
start = 0
while start < len(tokens):
end = start + chunk_size
chunk_tokens = tokens[start:end]
chunk_text = tokenizer.decode(chunk_tokens)
chunks.append(chunk_text)
start = end - chunk_overlap
return chunks
def chunk_by_sentences(
text: str,
max_chunk_size: int = 1000,
min_chunk_size: int = 100
) -> List[str]:
"""Chunk text by sentences, respecting size limits."""
import nltk
sentences = nltk.sent_tokenize(text)
chunks = []
current_chunk = []
current_size = 0
for sentence in sentences:
sentence_size = len(sentence)
if current_size + sentence_size > max_chunk_size and current_chunk:
chunks.append(" ".join(current_chunk))
current_chunk = []
current_size = 0
current_chunk.append(sentence)
current_size += sentence_size
if current_chunk:
chunks.append(" ".join(current_chunk))
return chunks
def chunk_by_semantic_sections(
text: str,
headers_pattern: str = r'^#{1,3}\s+.+$'
) -> List[Tuple[str, str]]:
"""Chunk markdown by headers, preserving hierarchy."""
lines = text.split('\n')
chunks = []
current_header = ""
current_content = []
for line in lines:
if re.match(headers_pattern, line, re.MULTILINE):
if current_content:
chunks.append((current_header, '\n'.join(current_content)))
current_header = line
current_content = []
else:
current_content.append(line)
if current_content:
chunks.append((current_header, '\n'.join(current_content)))
return chunks
def recursive_character_splitter(
text: str,
chunk_size: int = 1000,
chunk_overlap: int = 200,
separators: List[str] = None
) -> List[str]:
"""LangChain-style recursive splitter."""
separators = separators or ["\n\n", "\n", ". ", " ", ""]
def split_text(text: str, separators: List[str]) -> List[str]:
if not text:
return []
separator = separators[0]
remaining_separators = separators[1:]
if separator == "":
# Character-level split
return [text[i:i+chunk_size] for i in range(0, len(text), chunk_size - chunk_overlap)]
splits = text.split(separator)
chunks = []
current_chunk = []
current_length = 0
for split in splits:
split_length = len(split) + len(separator)
if current_length + split_length > chunk_size and current_chunk:
chunk_text = separator.join(current_chunk)
# Recursively split if still too large
if len(chunk_text) > chunk_size and remaining_separators:
chunks.extend(split_text(chunk_text, remaining_separators))
else:
chunks.append(chunk_text)
# Start new chunk with overlap
overlap_splits = []
overlap_length = 0
for s in reversed(current_chunk):
if overlap_length + len(s) <= chunk_overlap:
overlap_splits.insert(0, s)
overlap_length += len(s)
else:
break
current_chunk = overlap_splits
current_length = overlap_length
current_chunk.append(split)
current_length += split_length
if current_chunk:
chunks.append(separator.join(current_chunk))
return chunks
return split_text(text, separators)
```
### Template 4: Domain-Specific Embedding Pipeline
```python
class DomainEmbeddingPipeline:
"""Pipeline for domain-specific embeddings."""
def __init__(
self,
embedding_model: str = "text-embedding-3-small",
chunk_size: int = 512,
chunk_overlap: int = 50,
preprocessing_fn=None
):
self.embedding_model = embedding_model
self.chunk_size = chunk_size
self.chunk_overlap = chunk_overlap
self.preprocess = preprocessing_fn or self._default_preprocess
def _default_preprocess(self, text: str) -> str:
"""Default preprocessing."""
# Remove excessive whitespace
text = re.sub(r'\s+', ' ', text)
# Remove special characters
text = re.sub(r'[^\w\s.,!?-]', '', text)
return text.strip()
async def process_documents(
self,
documents: List[dict],
id_field: str = "id",
content_field: str = "content",
metadata_fields: List[str] = None
) -> List[dict]:
"""Process documents for vector storage."""
processed = []
for doc in documents:
content = doc[content_field]
doc_id = doc[id_field]
# Preprocess
cleaned = self.preprocess(content)
# Chunk
chunks = chunk_by_tokens(
cleaned,
self.chunk_size,
self.chunk_overlap
)
# Create embeddings
embeddings = get_embeddings(chunks, self.embedding_model)
# Create records
for i, (chunk, embedding) in enumerate(zip(chunks, embeddings)):
record = {
"id": f"{doc_id}_chunk_{i}",
"document_id": doc_id,
"chunk_index": i,
"text": chunk,
"embedding": embedding
}
# Add metadata
if metadata_fields:
for field in metadata_fields:
if field in doc:
record[field] = doc[field]
processed.append(record)
return processed
# Code-specific pipeline
class CodeEmbeddingPipeline:
"""Specialized pipeline for code embeddings."""
def __init__(self, model: str = "voyage-code-2"):
self.model = model
def chunk_code(self, code: str, language: str) -> List[dict]:
"""Chunk code by functions/classes."""
import tree_sitter
# Parse with tree-sitter
# Extract functions, classes, methods
# Return chunks with context
pass
def embed_with_context(self, chunk: str, context: str) -> List[float]:
"""Embed code with surrounding context."""
combined = f"Context: {context}\n\nCode:\n{chunk}"
return get_embedding(combined, model=self.model)
```
### Template 5: Embedding Quality Evaluation
```python
import numpy as np
from typing import List, Tuple
def evaluate_retrieval_quality(
queries: List[str],
relevant_docs: List[List[str]], # List of relevant doc IDs per query
retrieved_docs: List[List[str]], # List of retrieved doc IDs per query
k: int = 10
) -> dict:
"""Evaluate embedding quality for retrieval."""
def precision_at_k(relevant: set, retrieved: List[str], k: int) -> float:
retrieved_k = retrieved[:k]
relevant_retrieved = len(set(retrieved_k) & relevant)
return relevant_retrieved / k
def recall_at_k(relevant: set, retrieved: List[str], k: int) -> float:
retrieved_k = retrieved[:k]
relevant_retrieved = len(set(retrieved_k) & relevant)
return relevant_retrieved / len(relevant) if relevant else 0
def mrr(relevant: set, retrieved: List[str]) -> float:
for i, doc in enumerate(retrieved):
if doc in relevant:
return 1 / (i + 1)
return 0
def ndcg_at_k(relevant: set, retrieved: List[str], k: int) -> float:
dcg = sum(
1 / np.log2(i + 2) if doc in relevant else 0
for i, doc in enumerate(retrieved[:k])
)
ideal_dcg = sum(1 / np.log2(i + 2) for i in range(min(len(relevant), k)))
return dcg / ideal_dcg if ideal_dcg > 0 else 0
metrics = {
f"precision@{k}": [],
f"recall@{k}": [],
"mrr": [],
f"ndcg@{k}": []
}
for relevant, retrieved in zip(relevant_docs, retrieved_docs):
relevant_set = set(relevant)
metrics[f"precision@{k}"].append(precision_at_k(relevant_set, retrieved, k))
metrics[f"recall@{k}"].append(recall_at_k(relevant_set, retrieved, k))
metrics["mrr"].append(mrr(relevant_set, retrieved))
metrics[f"ndcg@{k}"].append(ndcg_at_k(relevant_set, retrieved, k))
return {name: np.mean(values) for name, values in metrics.items()}
def compute_embedding_similarity(
embeddings1: np.ndarray,
embeddings2: np.ndarray,
metric: str = "cosine"
) -> np.ndarray:
"""Compute similarity matrix between embedding sets."""
if metric == "cosine":
# Normalize
norm1 = embeddings1 / np.linalg.norm(embeddings1, axis=1, keepdims=True)
norm2 = embeddings2 / np.linalg.norm(embeddings2, axis=1, keepdims=True)
return norm1 @ norm2.T
elif metric == "euclidean":
from scipy.spatial.distance import cdist
return -cdist(embeddings1, embeddings2, metric='euclidean')
elif metric == "dot":
return embeddings1 @ embeddings2.T
```
## Best Practices
### Do's
- **Match model to use case** - Code vs prose vs multilingual
- **Chunk thoughtfully** - Preserve semantic boundaries
- **Normalize embeddings** - For cosine similarity
- **Batch requests** - More efficient than one-by-one
- **Cache embeddings** - Avoid recomputing
### Don'ts
- **Don't ignore token limits** - Truncation loses info
- **Don't mix embedding models** - Incompatible spaces
- **Don't skip preprocessing** - Garbage in, garbage out
- **Don't over-chunk** - Lose context
## Resources
- [OpenAI Embeddings](https://platform.openai.com/docs/guides/embeddings)
- [Sentence Transformers](https://www.sbert.net/)
- [MTEB Benchmark](https://huggingface.co/spaces/mteb/leaderboard)Related Skills
debugging-strategies
Transform debugging from frustrating guesswork into systematic problem-solving with proven strategies, powerful tools, and methodical approaches.
nft-standards
Master ERC-721 and ERC-1155 NFT standards, metadata best practices, and advanced NFT features.
nextjs-app-router-patterns
Comprehensive patterns for Next.js 14+ App Router architecture, Server Components, and modern full-stack React development.
new-rails-project
Create a new Rails project
networkx
NetworkX is a Python package for creating, manipulating, and analyzing complex networks and graphs.
network-engineer
Expert network engineer specializing in modern cloud networking, security architectures, and performance optimization.
nestjs-expert
You are an expert in Nest.js with deep knowledge of enterprise-grade Node.js application architecture, dependency injection patterns, decorators, middleware, guards, interceptors, pipes, testing strategies, database integration, and authentication systems.
nerdzao-elite
Senior Elite Software Engineer (15+) and Senior Product Designer. Full workflow with planning, architecture, TDD, clean code, and pixel-perfect UX validation.
nerdzao-elite-gemini-high
Modo Elite Coder + UX Pixel-Perfect otimizado especificamente para Gemini 3.1 Pro High. Workflow completo com foco em qualidade máxima e eficiência de tokens.
native-data-fetching
Use when implementing or debugging ANY network request, API call, or data fetching. Covers fetch API, React Query, SWR, error handling, caching, offline support, and Expo Router data loaders (useLoaderData).
n8n-workflow-patterns
Proven architectural patterns for building n8n workflows.
n8n-validation-expert
Expert guide for interpreting and fixing n8n validation errors.