About the role#
As a Performance Tools Intern, you will build the infrastructure required to profile and analyze our custom ML accelerators. You will work with hardware, compiler, firmware, and inference teams to create tools that help engineers understand system behavior and optimize performance. This role focuses on building the visibility needed to identify bottlenecks across CPUs, accelerators, and distributed workloads.
What you'll do#
- Build components for performance analysis and profiling infrastructure.
- Collect and analyze performance data, including hardware counters, execution traces, and memory behavior.
- Develop tools to trace host-side runtime activity, system behavior, and accelerator execution.
- Correlate performance events across CPUs, accelerators, storage, networking, and distributed workloads.
- Build visualization tools to help engineers identify bottlenecks and optimize models.
- Implement a data collection framework for hardware performance counters on a PCIe-based accelerator.
- Develop a user-space service for low-overhead tracing of accelerator activity.
- Design a correlated timeline view to visualize CPU API calls, driver submissions, PCIe transfers, and accelerator execution units.
- Create an analysis pass to detect memory access inefficiencies and PCIe bandwidth saturation.
What you'll need#
- Strong programming skills in C++ or Rust. Experience with Python is a plus.
- A solid understanding of computer architecture, including CPUs, GPUs or AI accelerators, memory hierarchies, and parallel programming.
- Experience or interest in low-level performance analysis, profiling, and optimization.
- Familiarity with tools such as Nsight, VTune, Xprof, or Perfetto is a plus.
- Experience or interest in operating systems, compilers, firmware, drivers, or other low-level systems software.
- A pursuit of degrees in Computer Science, Computer Engineering, or Electrical Engineering.
- Preferred: Experience developing performance analysis or debugging tools, working with ML accelerator architectures, or kernel-mode driver development for Linux or Windows.
Location & details#
- This is a full-time, paid internship.
- The role is based on-site in San Jose, California.
- Positions are available on a rolling basis.
About Etched
Etched operates in the computer hardware manufacturing industry. The company builds frontier inference clusters. Founded in 2022, it maintains its headquarters in San Jose, California. The organization is privately held and employs approximately 496 people.
How to get in at Etched
Securing an internship at Etched requires speed because early applicants often receive more attention before the candidate pool becomes crowded. Intern Insider sends an instant alert the moment a role matching your target is published anywhere, helping you apply among the first. Applying early helps you stand out before the pile grows. You can also improve your response rates by reaching out to the recruiters behind these roles directly to ask about the position or a potential referral. Intern Insider surfaces the specific recruiters at Etched so you can make that connection. Reaching out to the right person is often more effective than submitting an application into a queue.
