Sigrid Jin
AI & infrastructure engineer · Seoul / Vancouver
about
I build systems where machine learning meets infrastructure — LLM serving, embedding models, vector search, and the Kubernetes platforms underneath them.
I spend time with instructkr, an open community for Korean LLM research, and write about retrieval, embeddings, and systems on Medium.
work
Sionic AI Developer Advocate Engineer
Joined in 2023 as the fifth member, right after the company was founded. I work across the stack — RAG systems, embedding model serving, and the Kubernetes platforms (EKS, GKE, on-premise) they run on — and spend the rest of my time on talks, writing, and developer advocacy.
Instruct.KR Organizer
An open community for Korean LLM instruction-tuning and evaluation research.
University of British Columbia Computer Science
Studying computer science between Seoul and Vancouver.
projects
- agentjson — a parser that repairs broken JSON output in AI agent pipelines
- hello-ansible-skills — Ansible automation skills for Claude Code
- crisp-py — CRISP: clustering multi-vector representations for denoising and pruning
- eggroll-embedding-trainer — gradient-free NDCG optimization for retrieval with evolution strategies
- s3-vectors-rs — an unofficial Rust CLI for Amazon S3 Vectors
- k8s-hard-way — building a Kubernetes cluster from scratch on a MacBook