AI Software Engineering - Workloads Validation Intern

About the role

Job Details:

Job Description:

The Role and the Impact

As an AI Software Engineering Undergraduate Intern, you'll play a pivotal role in advancing compiler technologies for heterogeneous AI workloads. You'll work on projects that enhance the optimization and scheduling of AI kernels, targeting both Intel CPUs and GPUs, delivering substantial performance improvements and enabling cutting-edge research application. Through hands-on development, you'll contribute to innovations in compiler infrastructure that address real-world challenges in AI systems.

Business group

The Data Center Group (DCG) leads innovation in enterprise-grade technology, enabling Intel's contributions to AI, cloud computing, and edge solutions. As part of the Software and AI Pathfinding team, you'll collaborate on initiatives that shape the future of AI systems and drive Intel's leadership in data-centric applications.

Key Responsibilities

- Collaborate on the design and implementation of compiler optimization techniques, including algebraic rewrite systems and hierarchical scheduling.

- Integrate novel abstractions into MLIR-based compilation pipelines for AI workloads.

- Develop resource-aware models and apply constraint evaluation for optimized scheduling decisions.

- Prototype and verify equivalence checking methods leveraging probabilistic and symbolic approaches.

- Leverage Intel architecture features (AMX, AVX-512) to enhance compiler performance targeting CPUs and GPUs.

- Document and present technical findings, with opportunities for open-source contributions and publications.

Qualifications:

Minimum Qualifications

You must be an active student (able to maintain student status for at least 1 year) to qualify for this role.

- Bachelors degree or student status in Computer Science, Electrical Engineering, or related field with 0-1 years of hands-on experience in domain with demonstrated skills gained through internships, academic projects, coursework, or hands-on training OR a Masters degree in a relevant field with no prior professional experience.

- Advenced Python development skills within compiler ecosystems (experience with MLIR/LLVM advantageous).

- Academic standing in a graduate program with a focus on compilers, programming languages, or AI optimization.

Preferred Qualifications

- Familiarity with LLVM ecosystem tools.

- Hands-on experience with autotuning techniques or cost modeling for compiler performance optimization.

- Knowledge of tensor compiler frameworks like Triton or similar solutions.

- Exposure to probabilistic algorithms and SMT solvers (e.g., Z3).

- Prior publications or technical projects in compilers, program synthesis, or AI optimization.

- Experience in agentic AI solutions

Join us to contribute to transformative advancements in AI software engineering and grow your expertise in compiler technologies for heterogeneous systems.