System Design Task: Speaking-Scenario Practice Feature + MVP Auth API

You are designing a new speaking-scenario practice capability for a language-learning app. Users select real-life scenarios (e.g., ordering coffee, job interview), practice by speaking, and receive real-time feedback and scores. You will also design a small MVP authentication/profile API to support this feature.

Part A — Speaking-Scenario Practice Feature

1. Functional Scope:

   • Allow users to browse, search, and select scenarios.
   • Present prompts to the user.
   • Record user responses (speech).
   • Provide real-time feedback and a score on the user's response.
   • Track user progress through scenarios.
   • Enable pausing and resuming practice sessions.

2. Architecture:

   • Propose necessary services/components (e.g., content service, session orchestration, speech-to-text (ASR), LLM for feedback, scoring, personalization, analytics).
   • Define APIs for these services and the data flow between the client, services, and data stores.

3. Data Model:

   • Define schemas for Scenario, Turn/Prompt, UserSession, Attempt, and UserProgress.
   • Describe indexing strategies for efficient scenario discovery and session retrieval.

4. LLM/ASR Integration:

   • Select appropriate ASR and feedback-generation models.
   • Describe the prompt strategy for the LLM.
   • Define guardrails, latency targets, token/cost budgets, and fallback mechanisms for model failures or rate limiting.

5. Personalization:

   • Explain how to adapt scenario difficulty, vocabulary, or pace based on user proficiency.
   • Identify key signals and features to log for personalization.

6. Scale and Reliability:

   • Estimate QPS, concurrency, and storage growth.
   • Propose strategies for caching, streaming, and backpressure.
   • Address handling of timeouts, retries, idempotency, and partial failures.

7. Safety and Compliance:

   • Consider content moderation, PII handling, data retention policies, privacy controls, and localization.
   • Address offline or poor-network connectivity scenarios.

8. Experimentation and Metrics:

   • Define success metrics (e.g., learning outcomes, engagement).
   • Establish guardrail metrics.
   • Describe how to A/B test different prompt variations or feedback mechanisms.

Part B — MVP API: User Authentication and Profile Management

1. Endpoints:

   • Register (email + password)
   • Authenticate (login)
   • Retrieve Profile
   • Update Profile
   • Delete Account
   • Specify request/response shapes and appropriate HTTP status codes for each endpoint.

2. Authentication & Authorization (AuthN/Z):

   • Choose between session-based authentication and JWT.
   • Define token Time-To-Live (TTL) and refresh mechanisms.
   • Discuss secure cookie vs. header usage for token transmission.
   • Describe password hashing strategy (can be mocked) and account lifecycle management.

3. Storage:

   • Use in-memory data structures for users, sessions/tokens, and profiles.
   • Define primary keys and concurrency handling for these in-memory stores.

4. Non-Functionals:

   • Implement rate limiting.
   • Specify input validation rules.
   • Outline error handling strategies.
   • Ensure idempotency for Register and Delete Account operations.
   • Describe logging and minimal observability requirements.

5. Constraints:

   • Focus on MVP first, deferring complex edge cases.
   • AI tools are permitted.
   • Optimize for development velocity while maintaining clear, usable APIs.

6. Evaluation:

   • Explain the trade-offs made in your design decisions.
   • Justify your data modeling choices.
   • Discuss how this MVP API could evolve into a production-ready service.