Report a bug

What went wrong?

Gemini API & Google AI — AI Academy

AI Academy/Courses/Gemini API & Google AI

⌘K

UK Sign in Sign up

IntermediateИнженерияgemini

Gemini API & Google AI

Hands-on course on Google Gemini API: from first request to multimodal apps with vision, audio, and video. Gemini 2.0 Flash, Gemini 1.5 Pro, Live API, function calling, Search grounding, code execution, context caching, and Vertex AI enterprise deployment. Real code in Python and JavaScript.

5modules

14lessons

326 mintotal time

Developers integrating Gemini into productsaudience

Progress

0%0 / 14 lessons

Sign in to save your progress.

Module 1

Getting Started with Gemini API

API key, SDK install, your first generateContent call, model selection, and core generation parameters.

Your First Gemini API Call

Get an API key in AI Studio, install google-generativeai, make your first generate_content call, and parse the response structure.

Gemini Model Lineup and Choosing for the Task

Gemini 2.0 Flash vs 1.5 Pro vs 1.5 Flash vs Ultra: context, price, speed, and how to pick a model for the task and budget.

Generation Parameters and System Instructions

system_instruction, temperature, top_p, top_k, max_output_tokens, stop_sequences, and safety_settings via generation_config.

Module 2

Multimodal Capabilities

Vision, audio, and video as input: inline base64, File API, image reasoning, transcription, and video understanding.

Vision: Images as Input

Passing images inline (base64) and via the File API, multiple images, image reasoning, OCR.

Audio and Video: Transcription and Understanding

Audio transcription, video understanding (analyzing an hour-long clip/YouTube), working with timecodes and frames.

File API: Upload, Limits, Lifecycle

upload_file / list_files / delete_file, supported MIME types, size limits, ephemeral files (48h).

Module 3

Advanced Generation Features

Function calling, Google Search grounding, and built-in code execution — extending the model with external capabilities.

Function Calling: Tools for the Model

Function declarations, AUTO/ANY/NONE modes via ToolConfig, parallel calls, the multi-turn tool-use loop.

Search Grounding: Answers Backed by Google

Google Search grounding, grounding_metadata, source attribution, when it actually helps.

Code Execution: The Built-in Interpreter

Built-in Python sandbox, math/data tasks, reading executable_code and code_execution_result.

Module 4

Production Patterns

Streaming, context caching, and the Live API — what real apps under load need.

Streaming Responses

stream=True / streamGenerateContent, chunk handling, usage aggregation, streaming together with function calls.

Context Caching: Saving on Repeated Context

create/list/delete CachedContent, savings on large repeated prefixes, choosing TTL, when caching is justified.

Live API: Realtime Voice and Video

Live API over WebSocket, audio formats (16k in / 24k out PCM), session management, turn-taking, interruptions.

Module 5

Enterprise & Ecosystem

Vertex AI for enterprise and the broader Google AI ecosystem: SDKs, Gemma, embeddings.

Vertex AI: Enterprise Deployment of Gemini

Vertex AI vs AI Studio, service accounts, OAuth (the cloudcode-pa pattern), private endpoints, enterprise features.

The Google AI Ecosystem: SDKs, Gemma, Embeddings

AI Studio for prototyping, Google AI SDK vs Vertex SDK, MakerSuite, Gemma open models, embeddings (text-embedding-004).

AAI AcademyEN Start learningCatalog

Overview

AI Academy Progress Achievements Leaderboard Insights Notes Notifications Team API Pricing Settings

Learning

Tracks8 Courses51

Libraries

Prompts30 Workflows12 Tools22 Compare Blog

Updated · May 14, 2026