You will work across a wide range of ML problems, and the problems change as the products and the technology do.
Recent examples from the team
Utility modeling: Move beyond event prediction (whether a user will like another user, whether a user will report someone) to model the utility each event delivers to users and to the system, and optimize for the utility of the system as a whole.
Cold start and data scarcity: Build models that work when data is missing or thin — prototyping stages, privacy constraints, rare events — and design the loop that collects the data those models need next.
LLM and agentic systems: Leverage LLMs and agentic systems in Recommendation, Trust & Safety, and Profile Enhancement, both as products in their own right and as a way to raise the performance of existing models.
Generative model evaluation: Define evaluation metrics for generative models that connect to business outcomes and that the model can be optimized against.
Our engineers also publish selected technical work on the Hyperconnect Tech blog (written in Korean).
How to Set ML Objectives
How Hyperconnect Built an LLM Explanation Policy
LLM-as-a-Judge for Explanation Quality
What You'll Do
You own ML projects end-to-end. You work alongside other ML Engineers and carry projects from a rough goal through launch and measurement.
Shape the Problem: You are given the goal in broad terms. You make it concrete: what problem the project actually solves, what falls outside it, and what technical direction it takes.
Define What Success Means: Translate the goal into a metric a model can optimize, and define the online and offline evaluation criteria that determine whether the target metric improved.
Design the Model and the Data Behind It: Choose the modeling approach that fits the problem and the metric you are optimizing. Design how training data is collected, processed, and fed back, so that user behavior in production becomes the data that improves the next model.
Collaborate Across Teams: Work with engineers, product partners, and brand counterparts across Match Group's global offices — Seoul, Palo Alto, LA, Vancouver, Dallas, Tokyo. Keep partners outside your project team informed, and surface blockers early enough for the team to address them.
What We're Looking For
Project Ownership (most important): Demonstrated experience owning an ML project end-to-end, including planning and coordinating the work required to deliver a measurable result.
Depth in AI/ML: A solid foundation in AI/ML with deep knowledge in at least one domain, backed by project experience in which you applied it.
Problem Definition and Technical Judgment: You turn a business and technical goal into an ML problem that can be solved and measured, and you make reasonable technical choices when the information is incomplete.
Data Literacy: You use exploratory data analysis to find the statistical properties and patterns in the data, and let what you find shape how you approach the problem.
ML Frameworks and Tooling: Proficiency in Python and frameworks such as PyTorch, TensorFlow, or JAX, and the ability to pick up new AI/ML tools quickly as the field moves.
Cross-Functional Collaboration: Experience working on projects with people from other functions, and the ability to keep stakeholders outside the project aligned on its progress, key decisions, and risks.
Professional Working Proficiency in English: Able to conduct business meetings and participate in complex discussions without requiring assistance.
Fluent in Korean: Sophisticated professional interactions with native-level precision. Essential for cross-functional collaboration within the Seoul office.
Global Operational Readiness: Ability to accommodate flexible working hours for regular early morning syncs with US teams (ET/PT).
Nice to Have
Research Track Record: Publications at top-tier AI/ML venues (NeurIPS, ICLR, ICML, ACL, RecSys, KDD, CVPR) or awards in related competitions.
Problem Discovery and Initiative: Experience identifying a problem or research direction on your own, without one being handed to you, and carrying it through to a paper or a business result.
Experimentation and Causal Analysis: Designing A/B tests, defining target KPIs, and using causal analysis and SQL-based analysis to turn data into decisions.
ML Systems Engineering: Software engineering capability to build the AI/ML systems around your models, such as training pipelines and serving systems.
Leading Projects Across Functions: Experience leading a project that spans multiple functions, and working with stakeholders and leadership outside the project team.
Fluent in English: Sophisticated professional interactions with native-level precision and an understanding of cultural nuances.
Hiring Process
Work type: Full-time
Legal Entity: Hyperconnect / Brand: Match Group AI
Recruitment Process: Resume Review > Live Coding Interview > Hiring Manager Interview > Technical Interview > Technical Deep-dive & Cultural Alignment Interview > Offer
(* Please note that part of the interview process may be conducted in English)
For the document screening stage, only successful candidates will be individually notified of the results.
Application documents: A detailed, career-based English resume (PDF) in free format
This position is open to candidates seeking to transfer as Alternative Military Service Researchers (전문연구요원). For those under this program, service management will be conducted in accordance with the relevant military service exemption laws.
#LI-YN1
- 제출해 주신 내용 중 허위 사실이 있거나 관련법 상 근로제공에 결격사유가 있는 경우 채용이 취소될 수 있으며, 필요시 사전에 안내된 채용 절차 외에도 추가 전형 및 서류 확인이 진행될 수 있습니다.
- 국가보훈대상자는 관계 법령에 따라 우대하오니, 해당되시는 분께서는 지원 시 고지해주시고 채용 시 증빙서류를 제출해주시기 바랍니다.
- 하이퍼커넥트가 채용하는 포지션에 지원하는 경우, 개인정보 처리에 관하여서는 본 개인정보처리방침이 적용됩니다: https://career.hyperconnect.com/privacy
#HPCNT
