DeepSeek in Python

Working with DeepSeek in Python

James Chapman

AI Curriculum Lead, DataCamp

Coming up...

Course goals

Access DeepSeek models
Solve real-world tasks
Build your own chatbot! 🤖
Use Python code to do this!

Expected knowledge

Subsetting Python lists and dictionaries
Control flow → if, elif, else
Loops → for, while

Not expected to know

AI or machine learning

The DeepSeek logo.

AI company based in China
Trained V3 and R1 models ~30x cheaper than market-leaders

Two modes, one model

A chat icon.

Chat mode

Generates conversations
Use: chatbots

A reasoning icon.

Reasoning mode

Deeper reasoning
Use: data analysis, research, coding assistance

Accessing DeepSeek models

The DeepSeek logo.

Accessing DeepSeek models

The DeepSeek API crossed out.

Accessing DeepSeek models

The Together.ai API with some of the models it supports.

What is an API?

Application Programming Interface

A computer sending a request to an API.

What is an API?

Application Programming Interface

A waiter delivering orders and food between the kitchen and tables. This mimics how APIs operate between devices and systems.

What is an API?

Our device location is being sent to an API, which is then passed to a database. The API returns the response from the system, the weather forecast, back to our device.

DeepSeek models via APIs

A computer sending a request to the Together.ai API.

DeepSeek models via APIs

A computer sending a request to the Together.ai API, which contains the model and any data and model parameters.

Why use DeepSeek via an API?

→ Chatbot

Augmenting personal workflows and tasks

→ API

Provides flexibility for integrating AI into products, experiences, and processes
Interact via programming language

A computer interacting with the Together.ai API.

Let's practice!

Working with DeepSeek in Python