🔭 I’m currently working on Deep Learning Compiler at Intel.
😄Programming Languages:
-
Proficient in C++ including template metaprogramming
-
Experienced with Python, Java, x86 asm.
⛹️♂️My C++ skills are particularly valuable for DL compiler development, especially in:
- Implementing IR optimizations (e.g., loop transformations(tiling))
- Hardware-specific kernel generation (e.g., vectorization, operator fusion)
- Memory reuse optimizations in computational graphs.
- Some common data-flow optimization passes based on SSA, like global scope constant folding, LICM etc.
🍎Deep Learning Expertise:
Mainly in NLP Domain:
- Word Embeddings: word2vec
- Recurrent Architectures: RNN, LSTM
- Transformer Models: Transformer architecture
⚡ My homepage:https://bruce11111.github.io/ & www.isxxh.com