- Where
- Santa Clara, CA · Irvine, CA · Austin, TX · Westlake Village, CA · Chandler, AZ · Burlington, VT
- Level
- Intern
About Marvell
Marvell’s semiconductor solutions are the essential building blocks of the data infrastructure that connects our world. Across enterprise, cloud and AI, and carrier architectures, our innovative technology is enabling new possibilities.
At Marvell, you can affect the arc of individual lives, lift the trajectory of entire industries, and fuel the transformative potential of tomorrow. For those looking to make their mark on purposeful and enduring innovation, above and beyond fleeting trends, Marvell is a place to thrive, learn, and lead.
Your Team, Your Impact
The AI era has fundamentally changed what a semiconductor package is. A package is no longer a container for a chip — it is the architecture. As AI accelerators push beyond the reticle limit of any single die, the package has become the integration layer where compute silicon, high-bandwidth memory, optical I/O, and die-to-die interconnects are assembled into a unified system. At Marvell, the advanced packaging team is where that integration happens — for the custom XPUs, AI accelerators, and high-speed connectivity platforms that the world's largest hyperscalers depend on to scale their AI infrastructure.
Marvell's packaging work is production-qualified and shipping at hyperscale. The team developed an industry-first modular RDL interposer platform that enables multi-die AI accelerator designs 2.8x larger than the largest possible single-chip implementation — integrating over 1,390mm² of silicon alongside HBM3/3E memory stacks across six redistribution layers, with shorter die-to-die interconnects and supply chain flexibility that traditional silicon interposers cannot match. The same team is integrating Co-Packaged Optics directly into custom XPU packages, enabling XPU-to-XPU connections at distances 100x longer than electrical cabling with higher bandwidth density and lower power. These are not research prototypes. They are production platforms, qualified with major hyperscalers and ramping now.
The problems this team works on — signal integrity at 224G+ SerDes speeds across multi-die substrates, power delivery for XPU packages approaching 1,000W, thermal management for chiplet stacks with HBM, mechanical reliability across advanced 2.5D and 3D integration schemes — are among the most technically demanding in the semiconductor industry. They are also among the most directly connected to doctoral research in electrical engineering, materials science, and applied physics. Marvell's Ph.D. Intern Program places doctoral candidates inside these active packaging development efforts, working on problems that are inseparable from the research they are already pursuing. What you will take away is something no university lab can replicate: the experience of solving a packaging problem that ships inside the world's most advanced AI silicon.
What You Can Expect
As our Ph.D. Advanced Packaging & Physical Integration Intern, every day you will work on packaging design and analysis challenges tied directly to production AI accelerator and connectivity platforms. Specifically, you can expect to:
Perform signal integrity (SI) and power integrity (PI) analysis for high-speed interfaces — including 224G SerDes, die-to-die interconnects, and HBM interfaces — across multi-die package substrates and interposers
Develop and execute thermal and mechanical simulation models for advanced 2.5D and 3D chiplet packages, evaluating heat dissipation, warpage, and reliability under production conditions
Contribute to physical design and layout of package substrates, RDL interposers, and chiplet integration schemes for custom AI accelerator platforms
Collaborate with chip design, assembly, and test engineering teams to validate packaging assumptions against silicon measurements and production data
Use industry-standard EDA and simulation tools — including Ansys, Cadence Sigrity, or equivalent — to model and optimize package-level electrical, thermal, and mechanical performance
Present…