Student Research Project: GPU-Accelerated Data Processing with Velox and cuDF

IBM Research Zurich seeks an intern to design and implement novel GPU-accelerated data processing operators using Velox and cuDF, focusing on performance and abstraction.

This internship at IBM Research Zurich centers on enhancing GPU-accelerated data processing through the Velox framework and cuDF library. The intern will identify inefficiencies in current SQL execution patterns and develop new cuDF operators that better align with Velox's computational needs, balancing granularity, performance, and maintainability.

Responsibilities

• Examine the Velox architecture and its operator execution model.
• Investigate cuDF primitives and their application in SQL execution.
• Analyze SQL workloads and TPC-H execution traces to spot inefficiencies and missing abstractions.
• Identify recurring computation patterns and potential operator candidates.
• Explore operator abstraction levels, defining and evaluating alternative boundaries.
• Design and implement new cuDF operators, integrating them into Velox.
• Create unit and integration tests to verify correctness and robustness.
• Assess performance via the TPC-H benchmark, focusing on throughput, kernel efficiency, and memory behavior.
• Derive general operator design principles and prepare results for publication (stretch goal).

Requirements

• Strong proficiency in C++ and Python.
• Experience with benchmarking and performance profiling.
• Basic knowledge of databases and SQL.
• Ability to work independently and learn new concepts quickly.

Nice to have

• Experience with SQL engines, data processing systems, or GPU programming (CUDA, cuDF, RAPIDS, etc.).
• Familiarity with query execution, relational algebra, columnar data layouts, and memory management.
• Familiarity with query optimization, compiler techniques, or IR-based systems.
• Prior research experience or publications in systems, databases, or high-performance computing.

What the company offers

• Support for students to publish at top-tier conferences.
• Expectation that the output results in an open-source contribution.
• Join an open, multicultural research environment that values different perspectives.
• Support for flexible working arrangements.

About the company

IBM Research Zurich fosters a diverse and inclusive workplace, aiming to help all genders and backgrounds thrive professionally while maintaining a healthy work-life balance.

• Commitment to fostering diversity and inclusion in the workplace.
• Goal to help all genders and backgrounds thrive professionally while maintaining a healthy work–life balance.