About the role#
Our ML Compiler team builds the software that runs PyTorch models on Samsung’s compute-near-memory AI accelerator. This includes a PyTorch device and compiler backend, a Triton-based kernel compiler, and an LLVM back end. During this summer internship, you will own a project to make PyTorch models run faster on new memory-centric hardware, moving from research to merged code.
What you'll do#
- Design a search space for Triton kernel configurations including tile sizes, launch-grid mapping, on-chip buffer budgets, and load style.
- Build legality rules to prune configurations that will not compile or fit in on-chip memory.
- Develop a kernel cost model calibrated against our simulator and profiler to rank candidates.
- Build a benchmarking harness and a persistent tuning cache.
- Integrate the tuner with the torch.compile Inductor autotuning path and Triton’s autotuner.
- Extend tuning to data placement and export tuned costs for a placement pass.
- Measure kernel-level and end-to-end LLM inference gains.
- Share results in weekly syncs and a final presentation.
- Collaborate with PyTorch and Triton developers on upstream autotuning hooks.
- Hand off documented, tested code and complete other responsibilities as assigned.
What you'll need#
- Pursuing a master's or PhD in Computer Science, Computer Engineering, or Electrical Engineering, with at least one academic quarter or semester remaining.
- Strong Python and C++ skills.
- Experience writing or tuning Triton kernels or working with PyTorch compiler internals like torch.compile, FX, or Inductor.
- Grounding in compilers or performance engineering, including IR and passes, cost models, roofline analysis, and profiling.
- Preferred: Experience with auto-tuning or auto-sharding (e.g., TVM/Ansor, Halide, Kernel Tuner, Alpa, GSPMD) or search methods such as Bayesian optimization or ILP.
- Preferred: Experience with MLIR/LLVM back ends, scratchpad or near-memory architectures, LLM inference kernels, or existing open-source PyTorch/Triton contributions.
Location & details#
- This is a full-time, paid internship for Summer 2027.
- The role is based on-site at our San Jose, California headquarters.
- Relocation and housing stipends are available for eligible candidates.
About Samsung Semiconductor
Samsung Semiconductor operates as a subsidiary of Samsung Electronics. The company manufactures semiconductor components for device makers. It started operations in 1974. The organization employs more than 10,000 people.
How to get in at Samsung Semiconductor
Applying to Samsung Semiconductor early matters because recruiters review applications as they arrive. Intern Insider sends an instant alert the moment a role matching your target is published anywhere, so you apply among the first, early applicants get seen before the pile grows. You should aim to be in the first wave of candidates to ensure your resume reaches a human reader. Reaching out to the right person often helps you stand out from the general applicant pool. Intern Insider surfaces the recruiters behind the company's roles so you can reach out directly to ask about the role or a referral, which materially improves response rates. A brief, professional message to a recruiter can clarify your interest and help you learn more about the team.
