diff --git a/.gitignore b/.gitignore
index fad2fff..db50ed6 100644
--- a/.gitignore
+++ b/.gitignore
@@ -16,3 +16,6 @@ documents/
 
 # PDF test files
 backend/test_doc.pdf
+chroma_db/
+test_doc.pdf
+
diff --git a/backend/core/vector_store.py b/backend/core/vector_store.py
index ec4a733..05be863 100644
--- a/backend/core/vector_store.py
+++ b/backend/core/vector_store.py
@@ -1,3 +1,5 @@
+import numpy as np
+np.float_ = np.float64
 import chromadb
 from chromadb.config import Settings as ChromaSettings
 from loguru import logger
diff --git a/backend/engines/embedder.py b/backend/engines/embedder.py
new file mode 100644
index 0000000..1018d49
--- /dev/null
+++ b/backend/engines/embedder.py
@@ -0,0 +1,171 @@
+"""
+Embedding + ChromaDB Storage Engine
+
+Uses Google Gemini's text-embedding-004 model.
+Stores chunks in ChromaDB with metadata for filtered retrieval.
+
+Collection naming: stockiq_{ticker_lowercase}
+  e.g. stockiq_reliance, stockiq_tcs
+"""
+
+import google.generativeai as genai
+from loguru import logger
+from tenacity import retry, stop_after_attempt, wait_exponential
+from ..core.vector_store import get_or_create_collection
+from ..core.config import settings
+
+genai.configure(api_key=settings.GEMINI_API_KEY)
+
+@retry(stop=stop_after_attempt(3), wait=wait_exponential(multiplier=2, min=4, max=60))
+async def embed_text(text: str) -> list[float]:
+    """Embed a single text string using Gemini text-embedding-004 with fallback."""
+    for model_name in ["models/text-embedding-004", "models/gemini-embedding-2"]:
+        try:
+            result = genai.embed_content(
+                model=model_name,
+                content=text,
+                task_type="retrieval_document"
+            )
+            return result["embedding"]
+        except Exception as e:
+            err_msg = str(e).lower()
+            if "not found" in err_msg or "404" in err_msg or "not supported" in err_msg:
+                logger.warning(f"Embedding model {model_name} not found or unsupported, trying fallback...")
+                continue
+            logger.error(f"Embedding failed for model {model_name}: {e}")
+            raise
+    raise Exception("All embedding models failed.")
+
+async def store_chunks_in_vector_db(
+    ticker: str,
+    document_id: int,
+    document_type: str,
+    fiscal_year: int | None,
+    chunks: list
+) -> int:
+    """
+    Embed all chunks and store in ChromaDB.
+    Returns number of chunks stored.
+    """
+    collection_name = f"stockiq_{ticker.lower()}"
+    collection = get_or_create_collection(collection_name)
+
+    stored_count = 0
+
+    for chunk in chunks:
+        try:
+            chroma_id = f"{ticker}_{document_id}_{chunk.chunk_index if hasattr(chunk, 'chunk_index') else chunk.get('chunk_index', 0)}"
+            content = chunk.content if hasattr(chunk, 'content') else chunk.get('content', '')
+            section_type = chunk.section_type if hasattr(chunk, 'section_type') else chunk.get('section_type', 'GENERAL')
+
+            embedding = await embed_text(content)
+
+            collection.add(
+                ids=[chroma_id],
+                embeddings=[embedding],
+                documents=[content],
+                metadatas=[{
+                    "ticker": ticker,
+                    "document_id": document_id,
+                    "document_type": document_type,
+                    "fiscal_year": fiscal_year or 0,
+                    "section_type": section_type,
+                    "section_title": chunk.section_title if hasattr(chunk, 'section_title') else chunk.get('section_title', ''),
+                    "chunk_index": chunk.chunk_index if hasattr(chunk, 'chunk_index') else chunk.get('chunk_index', 0),
+                    "token_estimate": chunk.token_estimate if hasattr(chunk, 'token_estimate') else chunk.get('token_estimate', 0),
+                }]
+            )
+            stored_count += 1
+
+        except Exception as e:
+            logger.warning(f"Failed to store chunk {chunk.chunk_index if hasattr(chunk, 'chunk_index') else chunk.get('chunk_index', '?') if isinstance(chunk, dict) else '?'}: {e}")
+
+    logger.info(f"Stored {stored_count}/{len(chunks)} chunks for {ticker} doc {document_id}")
+    return stored_count
+
+async def retrieve_chunks(
+    ticker: str,
+    query: str,
+    n_results: int = 8,
+    section_type_filter: str = None,
+    document_type_filter: str = None,
+    fiscal_year_filter: int = None
+) -> list[dict]:
+    """
+    Retrieve top-k relevant chunks for a query.
+    Apply optional filters for section type, document type, fiscal year.
+    """
+    collection_name = f"stockiq_{ticker.lower()}"
+
+    try:
+        collection = get_or_create_collection(collection_name)
+    except Exception:
+        return []
+
+    # Build where filter
+    where_conditions = []
+    if section_type_filter:
+        where_conditions.append({"section_type": {"$eq": section_type_filter}})
+    if document_type_filter:
+        where_conditions.append({"document_type": {"$eq": document_type_filter}})
+    if fiscal_year_filter:
+        where_conditions.append({"fiscal_year": {"$eq": fiscal_year_filter}})
+
+    where = None
+    if len(where_conditions) == 1:
+        where = where_conditions[0]
+    elif len(where_conditions) > 1:
+        where = {"$and": where_conditions}
+
+    # Embed the query
+    try:
+        query_embedding = await embed_text_query(query)
+    except Exception as e:
+        logger.error(f"Query embedding failed: {e}")
+        return []
+
+    count = collection.count()
+    if count == 0:
+        return []
+
+    kwargs = {
+        "query_embeddings": [query_embedding],
+        "n_results": min(n_results, count),
+        "include": ["documents", "metadatas", "distances"]
+    }
+    if where:
+        kwargs["where"] = where
+
+    results = collection.query(**kwargs)
+
+    chunks = []
+    if results and "documents" in results and results["documents"]:
+        for i, doc in enumerate(results["documents"][0]):
+            chunks.append({
+                "content": doc,
+                "metadata": results["metadatas"][0][i],
+                "distance": results["distances"][0][i],
+                "relevance_score": 1 - results["distances"][0][i]
+            })
+
+    return chunks
+
+@retry(stop=stop_after_attempt(3), wait=wait_exponential(multiplier=2, min=4, max=60))
+async def embed_text_query(text: str) -> list[float]:
+    """Embed a query string (different task_type from document embedding)."""
+    for model_name in ["models/text-embedding-004", "models/gemini-embedding-2"]:
+        try:
+            result = genai.embed_content(
+                model=model_name,
+                content=text,
+                task_type="retrieval_query"
+            )
+            return result["embedding"]
+        except Exception as e:
+            err_msg = str(e).lower()
+            if "not found" in err_msg or "404" in err_msg or "not supported" in err_msg:
+                logger.warning(f"Embedding model {model_name} not found or unsupported, trying fallback...")
+                continue
+            logger.error(f"Query embedding failed for model {model_name}: {e}")
+            raise
+    raise Exception("All query embedding models failed.")
diff --git a/backend/test_milestone5.py b/backend/test_milestone5.py
new file mode 100644
index 0000000..3b1f9f7
--- /dev/null
+++ b/backend/test_milestone5.py
@@ -0,0 +1,110 @@
+# backend/test_milestone5.py
+import os
+import shutil
+import urllib.request
+import asyncio
+from pathlib import Path
+from backend.engines.pdf_extractor import extract_pdf
+from backend.engines.semantic_chunker import chunk_document
+from backend.engines.embedder import store_chunks_in_vector_db, retrieve_chunks
+from backend.core.vector_store import get_chroma_client
+
+
+def prepare_test_pdf():
+    test_pdf = Path("test_doc.pdf")
+    if test_pdf.exists():
+        print(f"Using existing {test_pdf}")
+        return
+
+    # Check if test_doc.pdf is in the backend directory
+    backend_pdf = Path("backend/test_doc.pdf")
+    if backend_pdf.exists():
+        shutil.copy(backend_pdf, test_pdf)
+        print(f"Copied {backend_pdf} to {test_pdf}")
+        return
+
+    # Check local document directories
+    possible_dirs = [Path("documents/TCS"), Path("../documents/TCS"), Path("backend/documents/TCS")]
+    for d in possible_dirs:
+        if d.exists():
+            pdfs = list(d.glob("*.pdf"))
+            if pdfs:
+                # Use the smallest one for speed
+                pdfs.sort(key=lambda p: p.stat().st_size)
+                shutil.copy(pdfs[0], test_pdf)
+                print(f"Copied {pdfs[0]} to {test_pdf}")
+                return
+
+
+    # If no local PDF, download a sample
+    print("No local PDF found. Downloading a sample Reliance financial PDF...")
+    url = "https://www.rfil.co.in/pdf/financial/reliance-financial-2023.pdf"  # Fallback sample URL
+    try:
+        urllib.request.urlretrieve(url, test_pdf)
+        print(f"Downloaded sample to {test_pdf}")
+    except Exception as e:
+        print(f"Download failed: {e}")
+        # Try a simpler, reliable PDF if the above fails
+        simple_url = "https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf"
+        urllib.request.urlretrieve(simple_url, test_pdf)
+        print(f"Downloaded dummy PDF to {test_pdf}")
+
+async def run_milestone5_test():
+    # 1. Prepare PDF
+    prepare_test_pdf()
+
+    # 2. Extract PDF
+    result = extract_pdf("test_doc.pdf")
+    print(f"Pages extracted: {len(result['pages'])}")
+    print(f"Tables found: {len(result['tables'])}")
+
+    # 3. Chunk Document
+    chunks = chunk_document(result["pages"], "QUARTERLY_RESULTS")
+    print(f"Chunks created: {len(chunks)}")
+
+    # 4. Clean up any existing Chroma collection for TCS to ensure clean test
+    client = get_chroma_client()
+    try:
+        client.delete_collection("stockiq_tcs")
+        print("Cleared existing stockiq_tcs collection")
+    except Exception:
+        pass
+
+    # 5. Store chunks in Vector DB
+    stored = await store_chunks_in_vector_db(
+        ticker="TCS",
+        document_id=1,
+        document_type="QUARTERLY_RESULTS",
+        fiscal_year=2024,
+        chunks=chunks
+    )
+    print(f"Stored {stored} chunks in ChromaDB")
+
+    # 6. Retrieve and verify
+    query = "revenue growth and EBITDA margin trends"
+    print(f"Retrieving chunks for query: '{query}'")
+    results = await retrieve_chunks(
+        ticker="TCS",
+        query=query,
+        n_results=5
+    )
+
+    print(f"Retrieved {len(results)} chunks")
+    high_relevance_count = 0
+    for i, r in enumerate(results):
+        score = r['relevance_score']
+        section_type = r['metadata']['section_type']
+        preview = r['content'][:150].replace('\n', ' ')
+        print(f"  {i+1}. Section: {section_type} | Relevance Score: {score:.4f}")
+        print(f"     Preview: {preview}...")
+        if score > 0.4:
+            high_relevance_count += 1
+
+    # 7. Assertions
+    assert len(results) >= 3, f"Expected at least 3 retrieved chunks, got {len(results)}"
+    assert high_relevance_count >= 3, f"Expected at least 3 chunks with relevance score > 0.4, got {high_relevance_count}"
+
+    print("\nMilestone 5: ALL TESTS PASSED")
+
+if __name__ == "__main__":
+    asyncio.run(run_milestone5_test())