Making sense of PDFs with AI: Why we need both Sentence-BERT and FAISS

Ankush Thakur — Thu, 21 Aug 2025 15:52:21 +0000

When I first started exploring question-answering on PDFs, one thing confused me:
👉 Why do we use Sentence-BERT and FAISS together?
Can’t FAISS just create embeddings on its own?

Here’s the simple breakdown 👇

🔹 Sentence-BERT

It’s a neural network model.

Converts text into embeddings (vectors of numbers).

These vectors capture the meaning of the text.

Example: “dog” and “puppy” → end up with vectors close to each other.

🔹 FAISS (Facebook AI Similarity Search)

It doesn’t create embeddings.

Instead, it’s an efficient search engine for vectors.

Given a query vector, it finds the nearest neighbors (most similar chunks of text) super fast.

👉 Think of it like this:

Sentence-BERT = Translator (text → coordinates on a “map of meaning”)

FAISS = GPS (finds the closest points on that map in milliseconds)

💡 Together, they make semantic search possible:

Sentence-BERT gives us the “language of meaning” (embeddings)

FAISS makes searching through thousands or millions of embeddings lightning fast

Without Sentence-BERT → FAISS has nothing meaningful to compare.
Without FAISS → you can still compare embeddings, but it’s painfully slow at scale.

NextJs TypeScript setup with Prettier, Husky.

Ankush Thakur — Wed, 14 Jun 2023 20:06:07 +0000

Let's get started with the NextJs setup with TypeScript. The configuration will be scalable with added configurations.

Let's create our app first by running the following command:

npx create-next-app@latest

You'll be presented with options. You can select each according to your preferences.

What is your project named? nextJs-awesome-setup
Would you like to use TypeScript with this project? No / Yes
Would you like to use ESLint with this project? No / Yes
Would you like to use Tailwind CSS with this project? No / Yes
Would you like to use `src/` directory with this project? No / Yes
Use App Router (recommended)? No / Yes
Would you like to customise the default import alias? No / Yes

Till now every step was generic which you can also find from here.

Deleting package.lock.json file and node_modules folder.

Then run:

yarn

We have now converted our app to support yarn instead of npm. Which will be quite beneficial in the longer run.

Try running yarn dev to check if everything is working as expected.

yarn dev

Prettier configuration

yarn add -D prettier

This will install the required dependencies

touch .prettierrc

This will create configuration file for Prettier where you can define your own custom rules.
Some common rules:

{
  "semi": true,
  "trailingComma": "none",
  "singleQuote": true,
  "printWidth": 80
}

Paste these rules as is in the .prettierrc file we have just created.

Create .prettierignore to ignore certain files and folders

touch .prettierignore

Add node_modules and .next folders

Add this line in package.json under scripts

"format": "npx prettier --write ."

Now you can finally test if it works, run

yarn format

Hope, it worked.

Husky configuration

Some information about husky

Install husky

npm install husky --save-dev

Enable git hooks

npx husky install

To automatically have Git hooks enabled after install, edit package.json

npm pkg set scripts.prepare="husky install"

After running this you should have this in your package.json

{
  "scripts": {
    "prepare": "husky install" 
  }
}

Create a hook which gets triggered whenever you make a commit

npx husky add .husky/pre-commit "yarn format"
git add .husky/pre-commit

Now whenever you do a new commit, your code should format itself before commit.

Bonus point: You run any script here using && and it will execute whenever you make commit for example your can run test every time before you make commit by adding:

yarn format && yarn test

TLDR! This setup guide offers you code formatting and check before making a commit, so that you don't push anything that looks bad.

Happy Coding!

DEV Community: Ankush Thakur

Making sense of PDFs with AI: Why we need both Sentence-BERT and FAISS

NextJs TypeScript setup with Prettier, Husky.