What’s Your AI Tech Savviness Level?

Posted on by Brittany Young

Hey tech enthusiasts! Brittany here, your friendly neighborhood data science nerd who’s probably spent way too many late nights tinkering with AI models (and yes, my coffee consumption should probably be monitored at this point 😅).

With all the buzz around DeepSeek R1, LlamaIndex, and other AI innovations lately, I thought it would be fun to create a quick self-assessment to see where you stand in the AI tech knowledge spectrum. No judgment here—we’re all learning together!

Test Your AI Tech Savviness

Deepseek – Level 1: The Curious Beginner

  1. When you hear “DeepSeek R1,” you think:
    a) A submarine exploring the ocean depths
    b) Some kind of AI language model
    c) A research paper I should probably read
    d) Is that the sequel to R2-D2?

  2. Function calls in AI contexts are:
    a) When you phone an AI assistant
    b) Methods that allow AI models to execute specific operations
    c) Special coding frameworks
    d) I’ve genuinely never heard this term before

  3. What does “reasoning_content” refer to in DeepSeek R1?
    a) The model’s Chain of Thought process
    b) A performance metric
    c) A data validation technique
    d) A debugging tool

Deepseek - person looking confused at computer code showing AI model output

Deepseek – Level 2: The Informed Enthusiast

  1. LlamaIndex’s AgentWorkflow is primarily designed to:
    a) Organize your daily tasks
    b) Create a workflow for llamas (duh!)
    c) Enable agents to hand off control between different components
    d) Index large language models

  2. The relationship between DeepSeek R1 and function calls is:
    a) DeepSeek R1 natively supports function calls
    b) DeepSeek R1 doesn’t support function calls by default
    c) DeepSeek R1 only supports certain types of function calls
    d) Function calls are irrelevant to DeepSeek R1

  3. To implement a DeepSeek client that can read reasoning_content, you would:
    a) Use the standard OpenAI module
    b) Create a custom client by extending OpenAILike
    c) Use the built-in DeepSeek client from LlamaIndex
    d) This isn’t technically possible

Level 3: The Technical Practitioner

  1. When extending OpenAILike for DeepSeek R1, which methods would you need to override to capture reasoning_content?
    a) complete() and predict()
    b) _stream_chat() and _astream_chat()
    c) generate() and stream()
    d) parse_response() and format_output()

  2. In the streamed output from DeepSeek R1, reasoning_content is found:
    a) In a separate API call
    b) In delta at the same level as content
    c) Only in the final response
    d) In the metadata header

  3. The main challenge in making DeepSeek R1 work with AgentWorkflow is:
    a) Its inability to support function calls
    b) Its incompatibility with Python
    c) The model size being too large
    d) Licensing restrictions

diagram showing AI system architecture with DeepSeek and LlamaIndex components

Level 4: The AI Integration Expert

  1. To enable function calls with DeepSeek R1, one approach is to:
    a) Wait for official support from DeepSeek
    b) Build a wrapper that parses the text output into structured calls
    c) Use a different model entirely
    d) Add a prompt engineer to your team

  2. When implementing a custom function calling system, the main consideration is:
    a) Making sure the model understands the function documentation
    b) Optimizing for speed
    c) Ensuring backward compatibility
    d) Minimizing token usage

  3. The best way to extract structured data from DeepSeek R1’s output is:
    a) Regular expressions
    b) Parsing JSON objects from the text
    c) Using a secondary AI to interpret the output
    d) This isn’t possible with DeepSeek R1

Score Yourself!

1-4 correct: AI Curious – You’re just starting your journey. No worries, we all begin somewhere!

5-8 correct: AI Savvy – You’ve got a solid grasp of the concepts. You could definitely hold your own in a tech conversation!

9-12 correct: AI Wizard – Impressive! You might want to consider a career in AI if you haven’t already.

Look, I’ll be the first to admit that keeping up with all these AI developments can feel like trying to drink from a firehose (trust me, I’ve had moments where I’ve considered printing out documentation just to cuddle with it at night… kidding! Sort of.). The field is moving incredibly fast, and that’s what makes it so exciting!

If you scored lower than you expected, don’t sweat it! The fact that you’re here, engaging with this content, puts you ahead of 90% of people. And if you aced it? Well, maybe drop me a line—I might have some projects that could use your expertise!

What aspect of AI tech are you most curious about these days? Let me know in the comments, and I might just make my next quiz about that topic!

Related Posts