Why pre-screen operations research interns before the interview
Interns should not be screened as though they have five years of experience. What matters is whether the method fundamentals are there, whether they can get data into a usable state without hand-holding, and whether they say when they are stuck rather than producing something confident and wrong. A short screen asks what they did when a model refused to work.
What actually matters when screening Operations Research Intern candidates
- 01
Technical proficiency
Check fluency with LP, MIP and heuristic formulations: decision variables, constraints, objective functions, plus hands-on use of Gurobi, CPLEX, PuLP or OR-Tools in Python or R.
- 02
Systems and trade-offs
Probe coursework capstones, internships or competition entries: inventory policy models, vehicle routing, queueing or discrete-event simulations built in SimPy, Arena or AnyLogic with stated results.
- 03
Evidence and rigour
Test how they validated a model: sensitivity analysis, baseline comparison, sanity-checking infeasibility, and whether assumptions about demand or capacity were tested against real data.
- 04
Collaboration and communication
Assess how they translated model output for supply chain, logistics or finance stakeholders who do not read constraint matrices; look for dashboards, memos or recommendation decks.
Pre-screening questions to ask Operations Research Intern candidates
12 questions grouped by what they test. Ask the same set in every screen and score answers on a consistent scale, or send them as an async video screen and compare answers side by side.
Has done the work
3 questions01Can you describe a project where you used operations research to solve a problem?
Listen forA project, including coursework, with the problem, method and result explained in their own words.
Projects described only by title, or contributions to group work left unclear.
02What is your understanding of operations research?
Listen forA practical explanation involving decisions under constraints, rather than a textbook definition.
Definitions recited without application, or the field confused with general data analysis.
03Can you describe using analysis to identify a problem and propose a solution?
Listen forA concrete example where their analysis changed what somebody decided or did.
Analysis described without any conclusion, or examples that came entirely from a lecturer.
Fundamentals understood
4 questions04Do you have experience with optimisation methods?
Listen forFormulation understood, including objective, constraints and what makes a problem hard to solve.
Solvers used without understanding the formulation, or infeasibility never encountered.
05Which statistical methods are you most comfortable with?
Listen forMethods they can explain including assumptions, with honesty about what they have not used.
Long lists of techniques, or assumptions behind common methods not understood.
06Are you familiar with predictive modelling and where it fits here?
Listen forA sensible distinction drawn between predicting an outcome and optimising an actual decision.
Prediction and optimisation treated as the same thing, or models judged on accuracy alone.
07How do you make sure your work is accurate?
Listen forChecks described concretely, such as validating against known cases or testing intermediate results.
Accuracy described as being careful, or results never checked against anything independent.
Works unaided
2 questions08What experience do you have with statistical software or programming?
Listen forEnough fluency to load, clean and analyse a dataset without step by step instruction.
Tools used only in guided exercises, or no experience outside a teaching environment.
09Have you worked with databases or larger datasets?
Listen forSome experience joining and filtering data, with an awareness of how messy real data is.
Only clean teaching datasets used, or no exposure to querying data directly.
Asks for help early
3 questions10Have you presented findings to a non-technical audience?
Listen forExplanation pitched to the audience, with the recommendation made clearly rather than implied.
Presentations built around method detail, or the conclusion left for the audience to find.
11How do you handle it when a piece of analysis is not working?
Listen forReasonable independent investigation followed by a specific, well-framed question to a supervisor.
Struggling silently for days, or asking for help before attempting anything.
12How comfortable are you working with a team on research?
Listen forWillingness to share work in progress and accept correction without taking it personally.
Preference for working entirely alone, or work only shared once it is finished.
How to score responses
Score every candidate on the same four criteria immediately after the screen. At this stage you are shortlisting for panel interviews, not making the final call.
Technical proficiency
35%5Formulates a scheduling or routing problem aloud, names solver settings like MIP gap and time limits, and explains dual values correctly.
Systems and trade-offs
25%5Describes a model they built end to end, including data sources, runtime, solution quality, and the decision it actually informed.
Evidence and rigour
25%5Distinguishes model error from data error, runs scenarios before trusting output, and cites a case where results contradicted their assumption.
Collaboration and communication
15%5Explains a solver recommendation in plain operational terms, anticipates the pushback on feasibility, and documents assumptions where others can audit them.
Interns apply in volume and most get a glance. A one-way video screen gives everyone the same eight questions.
Try it on HirevireScreening FAQ
Process basics
How long should a pre-screening round for this role take?
Ten minutes across eight questions, answered async. Interns apply in volume, and a short structured screen is the only fair way to review everyone rather than the first few applications.
How much should coursework count?
It counts. At this stage a well-explained university project shows more than a vague internship. What matters is whether they can describe what they built and why.
Evaluating answers
What is the strongest signal when screening interns?
What they did when something did not work. Interns who investigated and then asked a specific question will learn quickly. Those who submitted the broken result will need supervision.
What should I not expect at this level?
Production experience, stakeholder management or large scale deployment are all unrealistic here. Screen for fundamentals, curiosity and honesty, and treat anything beyond that as a genuine bonus.
Go deeper on this role
Sanat has been hiring since 2012 and watching the recruitment industry change up close ever since, and turned that screening process into Hirevire's video screening platform. LinkedIn
