ํ๋ก์ ํธ ์๊ฐ
We are seeking an experienced AI/ML Engineer to lead the design and delivery of a production-grade, multi-agent AI platform for enterprise use cases in ITSM and HR. This is a hands-on, technical leadership position that combines deep expertise in large language models (LLMs), retrieval-augmented generation (RAG), hybrid search/retrieval systems, and classical ML approaches such as clustering, anomaly detection, and root-cause analysis.
The successful candidate will own end-to-end architecture โ from system design and deployment to performance optimization โ partnering closely with Product, Platform, and Security teams to ship secure, reliable, and scalable features for multi-tenant environments.
๋น์ ์ ์๋ฌด
As an AI/ML Engineer you will be responsible for:
- Architect and deliver multi-agent orchestration systems with planning, tool-use, agent-to-agent protocols, and memory management, ensuring tenant isolation and governance
- Build high-reliability RAG pipelines including chunking, indexing, hybrid retrieval (BM25 + vector), query rewriting, re-ranking, and grounding with citations
- Develop online inference workflows for LLM and classical ML models, maintaining SLOs for latency, cost, and quality
- Design hybrid enterprise search solutions across structured and unstructured data (documents, tickets, HRIS/ITSM systems, logs, metrics)
- Implement and optimize embeddings, ANN indices, query planning, and feedback-driven re-ranking
- Develop and productionize classical ML models for anomaly detection, clustering, incident classification, and root-cause analysis
- Establish evaluation and safety frameworks, including automated regression testing, red-team testing, and compliance with SOC2/ISO and regional data policies
- Implement guardrails for prompt-injection, PII protection, RBAC, and content filtering
- Mentor engineers, lead design reviews, and collaborate with PMs and Design on roadmap priorities and success metrics
- Partner directly with enterprise clients on architecture reviews, scalability planning, and performance optimization
์๊ตฌ ์ฌํญ
- 5+ years of experience building production-grade ML/AI systems, with at least 3 years shipping LLM or RAG-based systems at scale
- 3+ years of experience leading distributed, multi-service architectures
- Expert proficiency in Python and Java, with strong foundations in algorithms, data structures, and distributed systems
- Proven delivery experience with RAG architectures using hybrid retrieval, vector databases (FAISS, Milvus, Pinecone, Weaviate), and Elasticsearch/OpenSearch
- Hands-on experience with agent frameworks such as LangGraph, LangChain, AutoGen, or CrewAI
- Proficiency in classical ML techniques including anomaly detection, clustering, and supervised classification
- Familiarity with MLOps and infrastructure tools such as Kubernetes, Docker, MLflow, W&B, Prometheus, and Grafana
- Experience building secure, multi-tenant systems on AWS, Azure, or GCP
- Integration experience with enterprise systems like ServiceNow, Jira, Workday, or SuccessFactors
- Excellent command of English, both written and spoken, sufficient for professional communication in international and technical environments.
- Strong collaboration skills and ability to engage with cross-functional teams (Product, Security, Data Engineering)
Nice to have:
- Experience with graph-based reasoning or topology for root-cause analysis (Neo4j, Neptune)
- Familiarity with retrieval safety, differential privacy, and policy enforcement
- Advanced evaluation knowledge (RAGAS, DeepEval, or LLM-as-judge frameworks)
- Expertise in large-scale vectorization, caching, and cost optimization strategies
- Open-source contributions to LLM, RAG, search, or MLOps frameworks
ํ์ ์ ๋ํ์ฌ
Team Up์์๋ ์ต๊ณ ์ ์ ๋ฌธ๊ฐ๋ค์ด ์๊ตญ์์ ๊ทผ๋ฌดํ๋ฉด์๋ ๊ธ๋ก๋ฒ ๊ธฐ์ ์์ ์๊ฒฉ์ผ๋ก ์ปค๋ฆฌ์ด๋ฅผ ์์ ์ ์๋๋ก ์ง์ํฉ๋๋ค. 2020๋ ๋ถํฐ 500๋ช ์ด ๋๋ ์ธ์ฌ๋ฅผ ๊ธ๋ก๋ฒ ๊ธฐ์ ๊ณผ ์ฐ๊ฒฐํ์ฌ ๊ตญ๊ฒฝ์ ๋๋๋๋ ๊ธฐํ๋ฅผ ์ฐฝ์ถํ๊ณ ์ง์ญ ์ฑ์ฅ์ ์ด์งํด ์์ต๋๋ค. ์กฐ์ง์์ ๋ ์ผ์ ํํธ๋์ญ์ผ๋ก ์์๋ ์ด ํํธ๋์ญ์ ์ด์ ์ฐ๊ฒฐ, ์ฑ์ฅ, ๊ทธ๋ฆฌ๊ณ ๋ ๋์ ๋ฏธ๋๋ผ๋ ๊ณต๋์ ๋น์ ์ ๋ฐํ์ผ๋ก 7๊ฐ๊ตญ์ผ๋ก ํ๋๋์์ต๋๋ค.
Team Up๊ณผ ํจ๊ปํ๋ ์๊ฒฉ ๊ทผ๋ฌด์ ์ด์ ๊ณผ ํน์
์ ๋ฌธ์ ์ผ๋ก ์ฑ์ฅํ๊ณ ์กด์ค๋ฐ๊ณ , ๋ณด์ดํ์ ๋ฐ๊ณ , ์์คํ๊ฒ ์ฌ๊ฒจ์ง ๋ค๊ณ ๋๋ผ๋ ๋ฐ ํ์ํ ๋ชจ๋ ๊ฒ
์ฌ๋ฌ ์์น · ์์ ์๊ฒฉ