{"$schema":"https://raw.githubusercontent.com/jsonresume/resume-schema/v1.0.0/schema.json","basics":{"name":"윤수환 Suhwan Yoon","label":"AI Engineer","summary":"Five years of AI/NLP engineering on financial data (filings, news, stocks), taking LLM agents, RAG, and document extraction from research through deployed APIs. Currently building search and multi-agent systems at Wanted Lab.","location":{"city":"Seoul, South Korea"},"profiles":[{"network":"LinkedIn","username":"수환-윤-72110a168","url":"https://www.linkedin.com/in/%EC%88%98%ED%99%98-%EC%9C%A4-72110a168"},{"network":"Career Mate","username":"wanted.co.kr/ai/agent/career-mate","url":"https://www.wanted.co.kr/ai/agent/career-mate"}]},"meta":{"canonical":"https://saywise.com/member8228","version":"v1.0.0","lastModified":"2026-10-27T05:59:14.580Z"},"x_saywise":{"handle":"member8228","pronouns":null,"availability":null,"oneLiner":"Five years of AI/NLP engineering on financial data (filings, news, stocks), taking LLM agents, RAG, and document extraction from research through deployed APIs. Currently building search and multi-agent systems at Wanted Lab.","profileUrl":"https://saywise.com/member8228","markdownUrl":"https://saywise.com/member8228/profile.md","pdfUrl":"https://saywise.com/member8228/resume.pdf","archetype":{"code":"HE-MX","name":"Junction Artisan","url":"https://saywise.com/sca/12827cd3de18","rarity":"One of the first Junction Artisans on Saywise"},"aiStack":[{"name":"LangChain","description":null,"url":"https://langchain.com","storyCount":0},{"name":"Claude Code","description":null,"url":"https://claude.com","storyCount":0},{"name":"Llama","description":null,"url":"https://llama.com","storyCount":0},{"name":"Hugging Face","description":null,"url":"https://huggingface.co","storyCount":0}]},"work":[{"name":"(주)원티드랩","position":"AI 기술팀 정규직","location":"Seoul, South Korea","startDate":"2025-11-01","summary":"Replaced the BM25 position search engine with a SPLADE + Dense + Lexical 3-leg hybrid search, tuned via A/B tests. Post-search bounce rate fell from 35% to 28% and search-driven application conversion rose from 6.7% to 8.4%. Also built the career mentor multi-agent service, where resume diagnosis raised user resume levels, and migrated legacy position agent code to LangChain and LangGraph."},{"name":"웨이커코리아 주식회사","position":"AI Engineer → AI Engineer Leader","location":"Seoul, South Korea","startDate":"2022-10-01","endDate":"2025-08-01","summary":"Built a multi-agent AI chat API that uses MCP tools grouped per agent, LLM orchestration, and a GuardRail module blocking sensitive identifiers. Developed a financial stock analysis agent using Plan-and-Act workflows, a Llama-Index QA agent that validates extracted SEC filing data, and a Kafka-based news pipeline with translation and summarization APIs. Earlier work covered SEC filing extraction with LLM and multimodal models, plus model compression for translation (distillation, pruning, quantization, ONNX)."},{"name":"(주)디앤아이파비스","position":"NLP Developer 정규직","location":"Seoul, South Korea","startDate":"2021-07-01","endDate":"2022-10-01","summary":"Built a BM25 patent document search engine on self-hosted Elasticsearch. Domain-tuned Sentence-BERT for problem and solution sentence matching, improving accuracy by 1.2% over vanilla BERT. Developed an LDA-based keyword extraction and clustering API for patent documents."},{"name":"네이버 주식회사","position":"인턴 연구원","location":"Seoul, South Korea","startDate":"2021-02-01","endDate":"2021-05-01","summary":"Researched training methods for ColBERT in information retrieval. Used Learning-to-Augment to generate hard negative samples and improved the Triplet Loss training approach."}],"education":[{"institution":"경희대학교 대학원","studyType":"컴퓨터공학과 석사 (자연언어처리)","startDate":"2019-03-01","endDate":"2021-02-01"},{"institution":"경희대학교","studyType":"컴퓨터공학과 학사","startDate":"2013-03-01","endDate":"2019-02-01"}],"projects":[{"name":"Career Mate 커리어 멘토 에이전트","description":"Six workflows and a multi-agent system covering mentoring, mentor recommendation, resume diagnosis, and nudge message generation. Resume diagnosis raised data quality for 11,626 cumulative users."},{"name":"번역 모델 경량화","description":"Built a half-depth MBART with knowledge distillation, then applied PTQ and attention-based pruning. Exported to ONNX as a static graph for serving."},{"name":"Question Difficulty Estimation (Applied Sciences)","description":"Co-first-authored a paper on estimating question difficulty with an attention model for question answering."}],"skills":[{"name":"LangChain"},{"name":"LangGraph"},{"name":"Llama-Index"},{"name":"Agentic AI"},{"name":"Multi-Agent"},{"name":"MCP"},{"name":"Prompt Engineering"},{"name":"GuardRail"},{"name":"Hybrid Search"},{"name":"SPLADE"},{"name":"BM25"},{"name":"OpenSearch"},{"name":"Elasticsearch"},{"name":"FAISS"},{"name":"Reranking"},{"name":"A/B Testing"},{"name":"PyTorch"},{"name":"Scikit-Learn"},{"name":"Named Entity Recognition"},{"name":"Topic Modeling"},{"name":"Knowledge Distillation"},{"name":"Model Pruning"},{"name":"Post Training Quantization"},{"name":"ONNX"},{"name":"Python"},{"name":"FastAPI"},{"name":"Flask"},{"name":"PostgreSQL"},{"name":"Redis"},{"name":"Apache Kafka"},{"name":"AWS"},{"name":"Docker"},{"name":"Jenkins"},{"name":"GitHub Actions"},{"name":"Git"}]}