Chew Jin Yang
My interests are in machine learning inference infrastructure, data engineering, and distributed systems.
I am interested in internships and engineering roles where I can work on infrastructure, data platforms, machine learning systems, distributed systems, backend engineering, or developer tools.
Current Activities
Hardware IDE
I am building a hardware IDE aimed at improving productivity for embedded system engineers. The goal is to make an IDE’s context representation natively connect with an agent that assists across the pipeline from datasheets and netlists to schematics, PCB CAD, firmware code, and product 3D CAD.
AutoResearch
I am studying and experimenting with Andrej Karpathy’s AutoResearch-style workflow, where agents propose changes, run experiments, evaluate results, and improve research or engineering loops through measurable feedback.
Machine Learning Inference Infrastructure
I am learning from and experimenting with inference infrastructure repositories such as FlashAttention, TileLang, SGLang, and DeepGEMM. My focus is understanding GPU kernels, memory movement, batching, scheduling, and high-performance model serving.
Distributed Scraping
I am working on distributed scraping systems to support my tuition fees and living expenses. This also helps me build practical understanding of networking, proxies, scheduling, retries, deduplication, data pipelines, and enterprise security.
LLM Backend Cost Reduction
I am exploring how to cluster Codex Exec and Claude -p workers as lower-cost LLM backend infrastructure for startup LLM platforms. The goal is to reduce API costs while still supporting world-class LLM-powered services. This involves complex session management, worker orchestration, queueing, reliability, and practical production serving.
Programming
I am learning C++23, CUTLASS, and Rust, mainly for machine learning infrastructure, inference optimization, and MLOps-related engineering.
LeetCode
I practice LeetCode and algorithmic problem solving to strengthen my fundamentals in data structures, algorithms, systems interviews, and backend engineering interviews.
Contact
Feel free to reach me at devve52@gmail.com.