Workflow Engine for Self-Help Menu | Dataford Interview Questions

Build a workflow engine to automate DoorDash's self-help menu: given stubbed Workflow (DAG of START/END/IS_LATE/FULL_REFUND nodes), MockDoorDashAPI (get_order, get_current_eta, issue_refund) and WorkflowEngine.execute_workflow(workflow, order_id), implement the stubs (get_start_node, from_txt, execute_workflow) so the provided test passes.

Asked in the onsite - AI Code Craft stage. Use AI tooling locally; prompt engineering is evaluated. Reported follow-ups include partial refunds for slightly late orders and investigating 10-20 second execution time.

Input and Output

Implement execute_workflow(workflow, order_id). workflow is a JSON object with definition, orders, and current_etas. Return true when a FULL_REFUND node is reached and false otherwise. A late order has current_eta > promised_eta.

Constraints

workflow["definition"] contains one node per nonempty line.
Node types are START, END, IS_LATE, and FULL_REFUND.
There is at most one START node.
Each valid workflow is a DAG.
Each order contains an integer promised_eta, and each current ETA is an integer.

Function Signature

def execute_workflow(workflow, order_id):

Your question is Workflow Engine for Self-Help Menu. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Input and Output

Constraints

workflow["definition"] contains one node per nonempty line.
Node types are START, END, IS_LATE, and FULL_REFUND.
There is at most one START node.
Each valid workflow is a DAG.
Each order contains an integer promised_eta, and each current ETA is an integer.

Function Signature

def execute_workflow(workflow, order_id):

Your question is Workflow Engine for Self-Help Menu. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.