Employee Vector Search with ChromaDB

A hands-on Python script that shows how to:

1. Create a vector collection in Chroma DB.
2. Embed rich employee profiles with the Sentence-Transformers model all-MiniLM-L6-v2.
3. Query the data three ways:
   • Semantic similarity (natural-language search).
   • Metadata filtering (SQL-like filters on fields such as department, experience, or location).
   • Combined searches (similarity + metadata for pinpoint accuracy).

It’s a quick reference for anyone learning modern vector databases, HNSW indexes, and hybrid filtering.

---

🔧 Requirements

Tool	Version (tested)
Python	3.9 or newer
chromadb	^0.5
sentence-transformers	^2.7

Tip: GPU is not required; CPU is fine for the small demo set.

---

🏃‍♂️ Quick Start

1. Clone the repo

git clone https://github.com/<your-username>/<your-repo>.git
cd <your-repo>

2. (Optional) Create & activate a virtual env

python -m venv venv
# macOS/Linux
source venv/bin/activate
# Windows
venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Run the script

python similarity_employeedata.py

• You will see console output similar to:

Collection created: employee_collection
Collection contents:
Number of documents: 15
=== Similarity Search Examples ===

1. Searching for Python developers:
Query: 'Python developer with web development experience'
  1. Alex Rodriguez (employee_10) - Distance: 0.0483
  2. John Doe (employee_1)       - Distance: 0.1037
  3. Matthew Garcia (employee_14) - Distance: 0.2126

---

🧩 How the Script Works (High Level)

Step	Code Section	What Happens
1	SentenceTransformerEmbeddingFunction	Loads all-MiniLM-L6-v2 for 384-dimensional embeddings.
2	chromadb.Client()	Starts an in-process Chroma server.
3	create_collection(...)	Creates an HNSW index (space="cosine") with the custom embedder.
4	Loop over employees	Generates natural-language blurbs per employee.
5	collection.add(...)	Adds IDs, documents, and rich metadata rows to the collection.
6	collection.query(...) & collection.get(...)	Performs six demo queries (semantic, filtered, hybrid).

---

✏️ Feel Free to Tweak

• employees list → Replace with your own data source (e.g., from a CSV, JSON, or database).
• query_text examples → Try custom search prompts like "DevOps manager with AWS" or "HR specialist in Boston".
• Metadata schema → Extend with fields like salary, skills (as lists), certifications, projects, etc.

---

🙌 Acknowledgements

• Chroma DB team for the blazing-fast vector store
• Sentence-Transformers by UKP Lab
• IBM Agentic AI Certification
• Example employees are fictitious; any resemblance to real persons is coincidental.