DeepSeek AI Masterclass: Build AI Agents & Automate Work

Explore how DeepSeek is built and why it's powerful, compare DeepSeek V3 and L1, and understand its advanced features and how it stacks up to ChatGPT models.

Discover how DeepSeek revolutionizes LM training by transferring knowledge from existing models and their responses, using a mixture of experts for faster, lower-cost, high-accuracy results.

Compare DeepSeek R1 and V3 for model classification, detailing R1's focus on complex problem solving and logical reasoning versus V3's versatile programming, multitasking, reinforcement learning, and mixture of experts.

DeepSea v3 is a fully open-source large language model for general tasks, delivering strong programming, coding, math, and problem-solving capabilities at low cost with fast, efficient performance.

Explore the open-source DeepSeek R1, a powerful reasoning model designed for logical inference, math problem solving, and real-time decision making, suitable for research and education tools.

Advance from the fundamentals of dopesick to building real ai agents with the latest models, including a Dopesick chatbot published on a website and an interactive content creator exercise.

Build real ai agents powered by the latest models by creating pipelines and chatbots with drag-and-drop blocks, connecting knowledge bases, and automations across tools like Google Drive and Google Calendar.

Build your first basic AI agent powered by DeepSeek to create a functional chatbot by linking input and output blocks with a knowledge base and system instructions.

Test the chatbot and verify pipeline execution using your knowledge base, inputs, and the Appsec model to deliver concise answers about AI, automation, Zapier, Notion, and business efficiency.

Export and customize your chatbot and automation from the vector shift, including chatbot, form, bulk jobs, and voicebot exports, and learn how to connect them to your website.

Export and customize a fully functional chatbot, then publish it on your website using embed code, with options to personalize name, avatar, styling, and conversation settings.

Create an AI-powered content creator from scratch using four inputs and a two-stage model that summarizes documents before generating structured content with DeepSeek, plus Slack integration.

Build an automation that collects a YouTube URL, downloads the transcript, summarizes it with a deep seq model, and saves the result to Google Docs for testing.

Advance to powerful automations with the latest deep sick models, building agents that handle complex tasks and include an ai language tutor, a company chatbot, and a research assistant.

Explore three practice approaches to building AI automations: watch first then build (recommended), try first then check, or build together, using provided files and step-by-step instructions.

Create an AI language tutor automation using DeepSeek, chat memory, and a knowledge base to adapt to user level, provide grammar and pronunciation feedback, quizzes, and web search for data.

Build a powerful company chatbot by integrating input, knowledge base, open source, chat memory, Google Web Search, Slack, Google Sheets, and tracking and exporting as a sophisticated automation.

Create a research assistant with a six-step automation that queries open source Wikipedia, saves inputs and responses to Google Sheets, and uses chat memory to streamline research.

Create an ai-powered business proposal generator that produces an introductory paragraph and three tailored bullet points from client name, example proposal, and notes. Automate proposals and save time.

Discover two methods to run deep seq locally, download the AI model to your computer, and work offline with privacy, security, cost efficiency, and customization.

Set up DeepSeek with LM Studio to run a deep sink model locally on your computer, download the distilled llama 8b version, and chat offline.

Connect DeepSeek to Ollama locally by downloading the 8B model, installing the app, and running a PowerShell command, then use Docker Desktop to access a web UI on localhost.

Explore DeepSeek use cases across text generation, translation, coding, data analysis, and custom gpts and agents, with previews of its capabilities.

Test how DeepSeek handles code by building Python games with pygame, including snake and pong, using Trinket IO or VS Code. Create a ten-question quiz with random order and scoring.

Explore how R1 and V3 solve logical problems, using the bridge-and-torch puzzle and a 100-prisoners example to analyze thinking processes, strategies, and model comparisons.

Explore creative writing with DeepSeek by generating SEO-friendly content on advanced Zapier automations that boost productivity, integrate with Notion and powered tools, and provide practical prompts and structure.

Learn to build an AI planner that schedules meetings, sends Gmail emails, and manages Google Calendar using Deepsea R1 and Deepsea chat, with HTTP request integration for a hands-on workflow.

Explore additional Deepsea concepts and learn how to run Deepsea on your mobile phone. Review the Deepsea API and privacy policy for using this application.

Download deep seq from the app store or google store, sign in with email or google, and start a chat to generate articles, mirroring the browser with history and uploads.

Understand DeepSeek's privacy policy: data collection (account details, input text, third-party data), data sharing, encryption, international transfers, usage to improve services and train models, and user rights.

Compare ipsec api pricing with other models by examining r1 and v3 input costs per million tokens, revealing ipsec's prices are far cheaper than gpt-4, gpt-4 turbo, and cloud 3.5.

DeepSeek is an open-source model with publicly available source code you can download, modify, and distribute. It offers transparent review, customizable use, no vendor lock-in, local privacy, and community-driven collaboration.

Create and manage api keys for the DeepSeek platform, connect them to chatbots in vector shift, and automate tasks across web and messaging apps with ai-powered api integrations.

Use the together AI platform to run DeepSeek with data staying in the United States; sign in, open V3 or R1, and customize the system prompt.

Compare how v3 and r1 deep seek models handle reasoning with chain-of-thought. See how the digit-sum task 1 to 1000 and a mislabeling puzzle reveal contrasting thinking processes.

Compare deepseek r1 and chatgpt o1 on thinking process, speed, and problem solving; differences are minor, so test them to decide between fast o1 or more structured, recall-driven deepseek r1.

The lecture compares DeepSeek, Claude, and ChatGPT across coding tasks, logic puzzles, SEO writing, and stock analysis, highlighting speed, code quality, and problem-solving approaches.

Compare deep seq v3 and cloud 3.5 sonnet to assess coding strength, broad knowledge, and reasoning, noting deep seq offers detailed free responses while cloud is faster but limited.

Celebrates completing the course and mastering deep sink to build powerful AI agents, with automations that save time, and invites learners to return for updates.