Intern, Thermal Modeling & Simulation Engineer (AI Memory / HBM)

What You’ll Learn

• Project Overview: Every new generation of AI accelerator pushes more power through stacked memory, and temperature is now one of the limits on how fast and how dense HBM can be. Our lab builds compact thermal models that turn workload-driven power maps of a 3D-stacked HBM into transient temperature maps in minutes rather than the hours a full CFD run takes, while staying close to CFD accuracy. As an intern you will own one piece of this platform end to end: from the physics and the CFD reference to a validated model that architecture, power, and package teams use to make design decisions.
• Skills You’ll Learn:
  • Heat conduction in 3D-stacked die: TSV arrays, micro-bumps, underfill, and anisotropic effective properties
  • Compact and reduced-order thermal modeling
  • CFD modeling and validation with Simcenter Flotherm
  • Scientific computing at scale: sparse solvers, preconditioning, and fast Python/MATLAB code
  • Power-thermal co-simulation and thermal-aware design of AI memory systems

What You’ll Do

Samsung’s Power & Thermal Lab is seeking a thermal engineering intern to help build and calibrate models of HBM stacks and package.

We are a small, fast‑moving team where direct ownership and minimal busywork empower every member to tackle the same real‑world problems. Working with authentic data, real partners, and production‑grade results, our contributions flow straight into platforms used across the company.

As an intern, the models you build feed design discussions for future HBM products. You will work on real projects with tangible results, receive mentorship, and help shape the scope of work in a collaborative environment where hardware, software, and thermal experts work side‑by‑side on next‑generation AI memory technologies.

Location: Daily onsite presence at our San Jose headquarters in alignment with our Flexible Work policy

Reports to: SVP, R&D (Power & Thermal Lab)

• Build and calibrate CFD models (Simcenter Flotherm) of HBM stacks and packages under realistic cooling conditions, and use them as the reference for fast models.
• Extend our compact thermal model to new stack heights, die configurations, and base-die power, and quantify its accuracy against CFD in both steady state and transient.
• Develop homogenization methods that replace fine features (TSV arrays, micro-bump layers, underfill) with effective properties, and establish where they hold and where they break.
• Speed up thermal solvers by orders of magnitude through model-order reduction, sparse linear algebra and preconditioning, GPU acceleration, or machine-learning surrogates.
• Run transient simulations driven by real AI workload traces, locate spatiotemporal hotspots, and explore thermal-aware policies such as throttling, refresh control, and power-map scheduling.
• Evaluate advanced cooling and thermal-interface options (cold plates, two-phase cooling, TIMs, backside cooling) and translate them into boundary conditions and design guidance for HBM.
• Couple the thermal model with the lab's trace-driven power simulation engine so that power maps and temperature maps evolve together in time.
• Document methods and results clearly and present them to engineers and lab leadership.
• Complete other responsibilities as assigned.

What You Bring

• Pursuing a PhD (preferred) or Master's degree in Mechanical Engineering, Materials Science, Electrical Engineering, Applied Physics, or a related field.
• Must have at least 1 academic quarter/semester remaining
• Strong foundation in heat transfer, especially transient conduction in multi-layer solids and thermal contact resistance.
• Hands-on experience with numerical methods (finite-volume, finite-element, or finite-difference) and scientific programming in Python and/or MATLAB.
• Ability to judge simulation results critically (mesh convergence, boundary conditions, energy balance) and to explain discrepancies between models.
• Experience with Simcenter Flotherm, Ansys Icepak/Fluent, or COMSOL, preferred
• Experience with compact or reduced-order thermal models, model-order reduction, or ML surrogates for physical simulation, preferred
• Background in electronics packaging, 3D-IC / chiplet thermal management, or advanced cooling, preferred
• Familiarity with HPC, sparse linear solvers, or GPU programming, preferred
• You’re inclusive, adapting your style to the situation and diverse global norms of our people.
• An avid learner, you approach challenges with curiosity and resilience, seeking data to help build understanding.
• You’re collaborative, building relationships, humbly offering support and openly welcoming approaches.
• Innovative and creative, you proactively explore new ideas and adapt quickly to change.