CV

Education

• Incoming thesis master’s student, Computer Science, University of Toronto, 2027 (expected)
  • Supervisor: Prof. Xujie Si
  • Research direction: agentic kernel generation and optimization, machine learning for code verification, and deep learning interpretability.
• B.Sc. in Mathematics and Computer Science, McGill University, 2021–2027
  • GPA: 3.78/4.00

Industry experience

• May 2025 – August 2026: Machine Learning System Engineer Intern
  • Qualcomm, Toronto, Canada
  • Maintained machine learning infrastructure and automated testing pipelines supporting model deployment.
  • Applied graph optimization and ONNX model compilation to computer vision workloads on Snapdragon hardware.

Research experience

• 2025–2026: Research Collaborations, Prof. Xujie Si’s Lab (ICLR / NeurIPS 2026 projects)
  • LogicXGNN: contributed to the design and testing of graph-hashing-based predicate construction; evaluated predicate extraction and hyperparameter choices with fidelity comparisons against GLGExplainer and GraphTrail.
  • Neural Proposals, Symbolic Guarantees: tested molecular tile decomposition, ran small-scale experiments, and contributed to proof writing for a framework combining neural proposals with SMT-based graph assembly.
• May 2023 – August 2023: Undergraduate Research Assistant, McGill University
  • SAT-DIFF: developed a MaxSAT-based framework for computing minimal edit scripts between ordered trees; proposed two solver optimizations and evaluated SAT-DIFF against GumTree.
• May 2022 – May 2023: Student Researcher, Mila – Quebec AI Institute
  • Language models for type-error diagnosis: continued pre-training and fine-tuning of BERT models for type-error localization in OCaml programs; structural probes for syntactic structure.
  • Neural network verification and robustness: proposed a metric to estimate expansion of the verifiable input space under reduced neural specifications; evaluated zero-bias networks on computer vision benchmarks.

Selected project

• June 2026: Team Lead, Assistive Cyber Eyes — 2025 Snapdragon AI Innovation Challenge (China)
  • Led the team and developed an AI navigation-assistance prototype for blind users on Qualcomm SM8850 hardware.
  • Won Merit Award as one of 27 finalist teams from 1,000+ registered teams.

Skills

• Languages: Python, C++, C, Java, OCaml, SQL, R
• ML tools: PyTorch, ONNX, NumPy, pandas, TensorFlow
• Research methods: SAT/MaxSAT encodings, graph algorithms, predicate extraction, neural network verification, experimental evaluation

Honors and awards

• SPLASH Student Research Competition – Second Place (2023)
• McGill Undergraduate Student Research Award (2023)
• ENG WEEK Hackathon – First Place (2022)

Publications

Novice Type Error Diagnosis with Natural Language Models

Geng, C., Ye, H., Li, Y., Han, T., Pientka, B., & Si, X. (2022, November). Novice Type Error Diagnosis with Natural Language Models. In Asian Symposium on Programming Languages and Systems (pp. 196-214). Cham: Springer Nature Switzerland.

Scalar Invariant Networks with Zero Bias

Geng, C., Xu, X., Ye, H., & Si, X. Scalar Invariant Networks with Zero Bias. NeurIPS 2023 Workshop on Symmetry and Geometry in Neural Representations; PMLR, 2024.

An Optimal Structure-Aware Code Difference Framework with MaxSAT-Solver

Ye, H. (2023). An Optimal Structure-Aware Code Difference Framework with MaxSAT-Solver. SPLASH Companion’23, 43.

SAT-DIFF: A Tree Diffing Framework Using SAT Solving

Geng, C., Ye, H., Zhang, Y., Pientka, B., & Si, X. SAT-DIFF: A Tree Diffing Framework Using SAT Solving. arXiv preprint arXiv:2404.04731, 2024.

Learning Minimal Neural Specifications

Geng, C., Wang, Z., Ye, H., & Si, X. Learning Minimal Neural Specifications. NeuS 2025 (Oral).

LogicXGNN: Grounded Logical Rules for Explaining Graph Neural Networks

Geng, C., Zhao, Z., Wang, Z., Ye, H., Jiang, Y., & Si, X. LogicXGNN: Grounded Logical Rules for Explaining Graph Neural Networks. In International Conference on Learning Representations (ICLR), 2026.

Neural Proposals, Symbolic Guarantees: Neuro-Symbolic Graph Generative Modeling

Geng, C., Zhang, L., Zhang, M., Wang, Z., Ye, H., & Si, X. Neural Proposals, Symbolic Guarantees: Neuro-Symbolic Graph Generative Modeling. In Advances in Neural Information Processing Systems (NeurIPS), 2026. (accepted)