Software Engineering @ UWaterloo (class of 2028).
Rust SWE intern @ Huawei (Severless WASM Engine)
- CPU limiting of thread-level WASM sandboxes using native OS threads and cgroups v2. Existing implementation has large overhead, using green threads with 2 layers of CPU schedulers :(
- Monitoring sandbox network usage with
eBPF
kernel programming - Reducing busy-waiting during WASM
I/O
operations using epoch counting
Information Retrieval
@ Castorini, under Dr. Jimmy Lin
ragnarok
RAG framework, http://arxiv.org/abs/2406.16828rank_llm
reranking LLMs
llama-index
: developed the rankLLM rerank
package, working on the QueryPipeline
feature
I enjoy rock-climbing in my spare time (stereotypical ik)
📫 Email: [email protected] / [email protected]