Architecture Research Intern (Predictive Modelling & Simulations), 12-months

Who You Are

• You’re fascinated by uncertainty and enjoy understanding how complex systems behave over time.
• You see queues everywhere and naturally think in terms of arrivals, service, congestion, and waiting.
• You naturally think in terms of probabilities, outcomes, and trade-offs.
• You ask “what if we changed this?” and know how to build a model that answers it rigorously.
• You’re comfortable working with incomplete information and building models that help make sense of it.
• You can balance model fidelity against computational cost and decision deadlines.
• You communicate technical concepts clearly and can explain uncertainty without creating confusion.
• You’re analytical, curious, and motivated by solving real-world problems.
• You enjoy collaborating and know the best solutions come from combining different perspectives.
• You love turning research ideas into practical models that create measurable value.
• How You’ll Make an Impact
• Contribute to predictive models and simulation frameworks that support queue management and estimated wait-time forecasting.
• Apply queueing theory to model customer flow, congestion, and waiting (for example multi-server queues, priority queues, queues with abandonment and retrials, and queueing networks) and use it to derive fast, interpretable wait-time estimates.
• Build simulation models (discrete-event, agent-based, and/or system dynamics) to evaluate how proposed business changes affect wait times, service levels, abandonment, and overall customer experience before they are implemented.
• Compare analytical queueing results against detailed simulation and real operational data to determine where theory holds, where it breaks down, and which approach best suits each decision.
• Formulate and analyze continuous-time models of system behavior using ordinary and stochastic differential equations (for example fluid and diffusion approximations of queues, including heavy-traffic limits based on reflected Brownian motion), and validate them against detailed simulation and real operational data.
• Design and apply time-constrained optimization algorithms (such as simulation-based optimization, parameter tuning, and anytime or heuristic methods) that deliver good decisions within real operational time limits.
• Design, test, and refine models that capture uncertainty, customer behavior, and system dynamics.
• Apply probabilistic and statistical techniques to evaluate performance, calibrate and validate models, and improve predictive accuracy.
• Run scenario, sensitivity, and what-if analyses that help business leaders understand bottlenecks, trade-offs, and future outcomes.
• Analyze results and communicate key insights, assumptions, limitations, and recommendations to stakeholders.
• Collaborate with researchers, engineers, and business teams to turn theory into practical, deployable solutions.
• Explore innovative approaches that improve decision-making across AMA’s products, services, and operations.
• What You Bring to the Table
• You’re currently pursuing, or have completed, an MSc or PhD in Mathematics, Applied Mathematics, Physics, Statistics, Operations Research, Computational Science, or a closely related quantitative discipline.
• You have in-depth knowledge and hands-on experience of queueing theory, including Markovian and general queues (M/M/c, M/G/1, G/G/c), Little’s Law, queues with abandonment and priorities, queueing networks, transient and steady-state analysis, and heavy-traffic and diffusion approximations. You can fit queueing models to real data and judge when their assumptions hold.
• You have in-depth knowledge of simulation science, including discrete-event and/or agent-based simulation, Monte Carlo methods, variance reduction, input modelling, model verification and validation, and experimental design for simulation studies.
• You have a strong foundation in differential equations (ordinary, partial, and/or stochastic), including Brownian motion and diffusion processes, Fokker-Planck equations, and numerical methods for SDEs, and can apply them to model the dynamics of real systems.
• You have working knowledge of optimization, particularly time-constrained or real-time settings: simulation optimization, stochastic and dynamic programming, metaheuristics, or anytime algorithms.
• You have a strong foundation in probability and stochastic processes, with experience in one or more of the following:
• Markov chains and Markov decision processes
• Renewal and point processes (for example Poisson and non-homogeneous Poisson arrival processes)
• Statistical mechanics
• Stochastic differential equations
• You have strong statistical judgment and understand the nuances of uncertainty, including the distinction between prediction intervals and confidence intervals.
• You’re proficient in Python and can independently translate models and research concepts into reliable, well-tested code (for example NumPy, SciPy, SimPy, and optimization libraries).
• Belonging Matters Here AMA is committed to providing an inclusive, respectful, and accessible environment. If you require an accommodation at any time during our recruitment process, please contact [email protected]. We’ll work with you to feel supported throughout your journey with us.
• WORK MODEL:
• Remote
• We thank all applicants for their interest; however, only those selected for an interview will be contacted.
• Belonging Matters Here
• AMA is committed to providing an inclusive, respectful, and accessible environment. If you require an accommodation at any time during our recruitment process, please contact [email protected]. We’ll work with you to feel supported throughout your journey with us.