Yes. Every demonstration is designed to be rebuilt alongside the instructor rather than watched passively. You write the same code, run it against the same inputs, and see the same outputs at each stage. Keeping an editor and terminal open while you watch will give you the most from the Specialization.

DeepSeek is a family of large language models built by the AI company of the same name, offered both as a hosted API and as downloadable open weight models you can run on your own hardware. The family includes general chat models and reasoning models that work through problems step by step before answering. For developers, the practical significance is that the same application can be pointed at a hosted endpoint or at a locally served model without rewriting the application logic.

The main structural difference is that DeepSeek publishes open weights for several of its models, so you can run them yourself rather than only calling a hosted endpoint. That matters when data cannot leave your environment, when you need to control versioning, or when hosted inference costs do not fit the workload. The application patterns in this Specialization, including structured output, function calling, retrieval, and agent loops, are largely portable across providers, so the skills transfer even where the API surface differs.

You need an API key, and the demonstrations are sized so that following along costs very little. Course 1 covers account setup, key generation, and how to keep keys out of source control. If you prefer not to use the hosted API at all, Course 3 covers running DeepSeek models locally, and readings explain which parts of the earlier courses can be completed that way.

Python throughout, with the DeepSeek API, Pydantic for response validation, a vector store for embeddings and retrieval, SQLite for the restricted database tool, Model Context Protocol for tool exposure, FastAPI for serving, and Docker for containerization. All of them are free or have free tiers sufficient for the coursework.

Chat models respond directly, while reasoning models work through intermediate steps before producing an answer, which helps on multi-step logic, code correctness, and structured extraction at the cost of latency and token use. Course 2 covers choosing between the two as a design decision, since routing every request to a reasoning model is usually the wrong default.

Yes. Course 3 covers MCP in practice: setting up a server, connecting DeepSeek to it, exposing and calling tools through it, and working with multiple tools across servers. MCP is treated as the standard interface for giving an agent tool access, alongside the direct function calling approach built in Course 2.

Function calling lets the model request that your code run a specific function with specific arguments, rather than describing what should happen in prose. It is what connects a language model to calculators, external APIs, databases, and any other system you expose. Course 2 covers defining tool schemas, executing calls safely, routing across multiple tools, and recovering when a tool fails.

This course is completely online, so there’s no need to show up to a classroom in person. You can access your lectures, readings and assignments anytime and anywhere via the web or your mobile device.

Yes! To get started, click the course card that interests you and enroll. You can enroll and complete the course to earn a shareable certificate. When you subscribe to a course that is part of a Specialization, you’re automatically subscribed to the full Specialization. Visit your learner dashboard to track your progress.

Yes. In select learning programs, you can apply for financial aid or a scholarship if you can’t afford the enrollment fee. If fin aid or scholarship is available for your learning program selection, you’ll find a link to apply on the description page.