← BNP Paribas Interview Insights
Start by clearly defining the Greeks and their role in measuring risk sensitivities, then focus on delta as the first-order price sensitivity to the underlying. For the portfolio delta, explain the aggregation method (sum of delta-weighted positions) and walk through a concrete example with a few instruments, showing the calculation step by step.
Pro tip: Emphasize that delta is not static—it changes with the underlying price (gamma) and time (charm), so portfolio delta must be dynamically managed. Also, mention that for software engineers, implementing Greeks calculations requires careful handling of numerical precision and market data updates.
Briefly list the main Greeks (delta, gamma, vega, theta, rho) and state that they measure sensitivity of option prices to various factors. Highlight that delta is the most fundamental.
Define delta as the rate of change of option price with respect to the underlying price. Mention that call deltas are positive (0 to 1) and put deltas are negative (-1 to 0), and that delta approximates the hedge ratio.
Explain that portfolio delta is the sum of the deltas of individual positions, each weighted by the quantity and contract multiplier. For example, delta_portfolio = Σ (quantity_i × delta_i × multiplier_i).
Choose a simple portfolio (e.g., 100 shares of stock, 10 call options with delta 0.6, 5 put options with delta -0.4) and calculate the total delta. Show the arithmetic clearly.
Interpret the result: e.g., a delta of 50 means the portfolio gains $50 for a $1 increase in the underlying. Mention that delta hedging involves offsetting this exposure and that delta changes over time.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.
Went through linearity, independence of errors, homoscedasticity, normality, no perfect multicollinearity.
Start by listing the standard assumptions of linear regression (linearity, independence, homoscedasticity, normality of errors, no multicollinearity, and no influential outliers). Then, for each assumption, describe a diagnostic method (e.g., residual plots, statistical tests, VIF) and briefly explain how you would address violations if found.
Pro tip: Emphasize that in practice, some assumptions (like normality) are less critical for large samples due to the Central Limit Theorem, while others (like independence) are non-negotiable. This shows you understand the practical implications and can prioritize.
Clearly state the six key assumptions: linearity, independence of errors, homoscedasticity, normality of errors, no multicollinearity, and no influential outliers.
For linearity, use residual vs. fitted plots or component+residual plots. For independence, use Durbin-Watson test or plot residuals vs. order.
For homoscedasticity, use residual vs. fitted plots or Breusch-Pagan test. For normality, use Q-Q plots or Shapiro-Wilk test.
For multicollinearity, compute VIF or condition index. For outliers, use leverage (hat values), Cook's distance, or studentized residuals.
Briefly mention potential fixes: transformations, robust standard errors, removing outliers, or using regularized regression.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.
These came fast and without much warning between resume questions.
Treat the resume discussion as a two-way conversation: when the interviewer asks about a project or skill, be ready to explain the underlying algorithm or mathematical reasoning. Clarify the problem, state assumptions, and walk through your thought process step by step, even for short questions. If you don't know, reason aloud and connect to known concepts rather than guessing.
Pro tip: Interviewers at BNP Paribas often care more about how you think than the final answer. Verbalize your reasoning, check edge cases, and ask clarifying questions—this shows the analytical rigor expected in a regulated financial environment.
Repeat the question in your own words and ask clarifying questions about input size, constraints, or expected output. This ensures you understand the problem and buys time to think.
Connect the question to a similar problem you solved in a past project or coursework. Briefly mention the context and the approach you used, then adapt it to the current question.
State the algorithm or mathematical method you plan to use, including time and space complexity if relevant. Keep it concise but show you have a structured plan.
Use a small example to demonstrate your approach step by step. This makes your reasoning concrete and helps catch mistakes early.
Mention potential edge cases (e.g., empty input, large numbers) and how you would handle them. If time permits, suggest possible optimizations or alternative approaches.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.