Computer Science @ The University of Texas at Dallas
Google Summer of Code 2026 — CERN ROOT / TMVA-SOFIE

AI Runtime Systems · ML Compiler Infrastructure · Performance Engineering

I am a Computer Science student at UT Dallas working on production-scale machine learning compiler and runtime infrastructure.

My current work focuses on C++ ML inference systems, ONNX model execution, runtime optimization, code generation, dynamic tensor shapes, and performance engineering.

Google Summer of Code 2026 Contributor

I contribute to ROOT's TMVA/SOFIE, a C++ inference and code-generation system for machine learning models.

10 upstream code reviews

My work spans compiler infrastructure, generated inference code, runtime execution, memory optimization, profiling, and performance engineering.

Extended SOFIE's model and runtime infrastructure for dynamic tensor dimensions, shape tensors, runtime allocation, and operator propagation.

Extended SOFIE's model and runtime infrastructure for dynamic tensor dimensions, shape tensors, runtime allocation, and operator propagation.

Added generated profiling infrastructure for analyzing inference execution and runtime performance.

Made SOFIE-generated C++ inference headers self-contained by emitting only the runtime helper implementations required by each model.

Improved ConvTranspose inference execution by reducing temporary memory usage and unnecessary memory allocation.

Improved SOFIE optimization and generated execution paths through compiler-level and operator-level optimizations.

• ONNX parser improvements and operator support
• Runtime broadcasting and tensor execution
• Session memory allocation
• Einsum BLAS optimization
• Constant broadcasting
• Modern C++ API safety improvements
• Testing and CI infrastructure
• Keras reproducibility and framework compatibility

Explore ROOT/TMVA-SOFIE →

C++ ROOT TMVA SOFIE ONNX Linux CMake Git

A production learning platform designed and developed to support large-scale online learning and assessment workflows.

94,000+ users

Engineering work includes:

• Django/Python backend infrastructure
• REST APIs
• Authentication and user data systems
• Database architecture
• Runtime request observability
• API rate limiting and security
• Production deployment and monitoring

Technologies:
Python Django SQL REST APIs JavaScript HTML/CSS

Visit APStudy →

C++ · Python · Java · SQL

ROOT · TMVA · SOFIE · ONNX · Alpaka

Linux · CMake · Docker · Git · REST APIs

• AI Runtime Systems
• Machine Learning Compiler Infrastructure
• GPU Runtime
• Performance Engineering
• Code Generation
• Memory Optimization
• Large-Scale AI Systems

The University of Texas at Dallas
B.S. Computer Science · Expected May 2030

Academic Excellence Scholarship (AES)

• GitHub: @Yongyanliu65
• Email: dal851569@utdallas.edu

Interested in AI runtime systems, ML compilers, high-performance inference, and production open-source infrastructure.