Scholar
Shangyin Tan
Google Scholar ID: NP2VIUUAAAAJ
PhD Student, UC Berkeley
Citations & Impact
Citations
199
H-index
8
i10-index
6
Publications
12
Co-authors
6
list available
Publications
Resume (English only)
Academic Achievements
- - Preprints (Work in Progress):
- - GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
- - Choix: Choice-based Learning in Jax
- - DSPy Assertions: Computational Constraints for Self-Refining Language Model Pipelines
- - Publications:
- - Programming Large Language Models with Algebraic Effect Handlers and the Selection Monad
- - LangProBe: a Language Programs Benchmark
- - ItyFuzz: Snapshot-Based Fuzzer for Smart Contract
- - Compiling Parallel Symbolic Execution with Continuations
- - INTENT: Interactive Tensor Transformation Synthesis
- - Towards Partially Evaluating Symbolic Interpreters for All
- - LLSC: A Parallel Symbolic Execution Compiler for LLVM IR
- - Compiling Symbolic Execution with Staging and Algebraic Effects
Research Experience
- - Student researcher at Google DeepMind (legacy Google Brain), collaborators: Dan Zheng, Ningning Xie, Gordon Plotkin
Education
- - Ph.D. student at UC Berkeley EECS, advisors: Koushik Sen and Matei Zaharia
- - Undergraduate at Purdue University, advisors: Guannan Wei and Tiark Rompf
Background
- Research interests center around agents, compound AI systems, and a little bit of programming languages.
Miscellany
- Hobbies include hiking/backpacking, trail running, and alpine skiing.
