Fireworks AI Quiz

Answers and explanations appear as you take the quiz.

1. What is Fireworks AI mainly known for?
   • Fast, low-cost inference serving for generative models
   • Training frontier models from scratch
   • Selling GPUs
   • Building self-driving cars
2. Why can developers often migrate to Fireworks with little code change?
   • It requires rewriting everything in Rust
   • It uses an OpenAI-compatible API
   • It only works in spreadsheets
   • It has no API
3. What is 'speculative decoding,' an optimization Fireworks uses?
   • Guessing the user's password
   • Encrypting model weights
   • A small draft model proposes tokens that the big model verifies in parallel
   • Randomly skipping words
4. Which kind of models does Fireworks primarily host?
   • Only its own secret models
   • Only image-editing plugins
   • Physical robots
   • Open-weight and custom models like Llama, Mixtral, and DeepSeek
5. What technique shrinks a model's memory and bandwidth needs to serve it faster?
   • Quantization
   • Recompilation of Windows
   • Adding more parameters
   • Disabling the GPU
6. Besides hosting, what additional capability does Fireworks offer?
   • Manufacturing semiconductors
   • Fine-tuning with LoRA adapters
   • Domain name registration
   • Email hosting
7. What is the purpose of continuous batching in a model-serving system?
   • Compress model weights into fewer bits
   • Convert text output into speech
   • Combine requests arriving at different times so the GPU stays productively occupied
   • Remove all requests that are still running
8. Which structured-output capability can Fireworks provide for an agent workflow?
   • Automatic GPU chip manufacturing
   • A guarantee that every tool call is safe
   • An image-only output format
   • JSON-shaped responses or function calls that applications can consume