AI Code Craft - Design a Code Generation and Review Pipeline. LeakCode

Question Details

Round 1 Coding / System Design

Problem

Design an internal tool that takes a natural language task description and outputs working code with an automated review step. The pipeline has three stages: (1) code generation, (2) static analysis, (3) test execution.

Part A - Code: Implement the pipeline orchestrator.

python
class CodeCraftPipeline:
    def run(
        self,
        task: str,
        language: str = "python"
    ) -> dict:

**Returns**:
        # {
        #   "code": str,
        #   "lint_issues": list[str],
        #   "test_results": {"passed": int, "failed": int},
        #   "status": "approved" | "needs_review"
        # }
        ...

    def generate_code(self, task: str, language: str) -> str: ...
    def lint(self, code: str) -> list[str]: ...
    def run_tests(self, code: str) -> dict: ...


**Example**:
pipeline = CodeCraftPipeline()
result = pipeline.run("write a function that reverses a linked list")
# result["code"]         -> "def reverse_list(head): ..."
# result["lint_issues"]  -> []
# result["test_results"] -> {"passed": 3, "failed": 0}
# result["status"]       -> "approved"

Follow-ups

How would you sandbox test execution to prevent malicious code from running?
What feedback loop would you add so lint and test failures are fed back into code regeneration?
How would you version and store generated code artifacts?
How do you evaluate the quality of generated code beyond passing tests?

Full Details

Round 1 Coding / System Design

Problem

Design an internal tool that takes a natural language task description and outputs working code with an automated review step. The pipeline has three stages: (1) code generation, (2) static analysis, (3) test execution.

Part A - Code: Implement the pipeline orchestrator.

python
class CodeCraftPipeline:
    def run(
        self,
        task: str,
        language: str = "python"
    ) -> dict:

**Returns**:
        # {
        #   "code": str,
        #   "lint_issues": list[str],
        #   "test_results": {"passed": int, "failed": int},
        #   "status": "approved" | "needs_review"
        # }
        ...

    def generate_code(self, task: str, language: str) -> str: ...
    def lint(self, code: str) -> list[str]: ...
    def run_tests(self, code: str) -> dict: ...


**Example**:
pipeline = CodeCraftPipeline()
result = pipeline.run("write a function that reverses a linked list")
# result["code"]         -> "def reverse_list(head): ..."
# result["lint_issues"]  -> []
# result["test_results"] -> {"passed": 3, "failed": 0}
# result["status"]       -> "approved"

Follow-ups

How would you sandbox test execution to prevent malicious code from running?
What feedback loop would you add so lint and test failures are fed back into code regeneration?
How would you version and store generated code artifacts?
How do you evaluate the quality of generated code beyond passing tests?

Free preview. Unlock all questions →

Topics

Coding Onsite

More from DoorDash

Reddit DoorDash Gets Rid Of LeetCode Interviews In Favor Of AI Technical Interview Reddit DoorDash Analytics Engineer Technical Interview Reddit Doordash phone screen yay or nay ? 1p3a DoorDash Senior Engineer Interview: Code Craft and System Design 1p3a Doordash Software Engineer Onsite Interview Experience and Insights