Subhiksha interview questions

2-Hour Coding Exercise – Multi-Agent Expense Monitoring & Analysis System Objective Build a multi-agent AI application that helps users: Monitor expenses Analyze spending patterns Generate actionable financial insights Scenario A fintech startup wants to create an AI-powered personal finance assistant. Users upload or provide expense data, and the system uses multiple AI agents to: Categorize expenses Detect anomalies or overspending Generate recommendations and summaries Expected Outcome Candidates should build 3 collaborating agents: Agent 1 – Expense Monitoring Agent Responsibility Read expense transactions Validate and normalize data Categorize expenses Agent 2 – Expense Analysis Agent Responsibility Analyze categorized expenses and identify: Highest spending category Spending trends Unusual transactions Daily/weekly totals Expected Insights Examples: “Food expenses increased by 40% compared to transport.” “You spent unusually high on dining this week.” “Average daily spend: ₹1,250.” Agent 3 – Recommendation Agent Responsibility Generate actionable recommendations based on insights. Example Recommendations “Reduce food delivery expenses by setting a weekly cap.” “Transportation costs are stable and within expected range.” “Subscriptions exceed 20% of discretionary spending.”