Machine Learning Engineer

Key Skills

Aartificial intelligence, API, Effective communication skills, Machine learning techniques, NeurIPS, Optimization, SQL

Job Description

[Recommendation]

By solving various recommendation problems included in the product, we provide a better experience to users and ultimately contribute to long-term sales increase. We are looking for people who can solve the following problems together. ( Squad Interview )

• Cold-start recommendation problem to provide a good experience to new users (systems that can identify user preferences with only a few-shot data, such as session-based recommendation, graph-based recommendation, and contextual bandit, and learning methods to improve recommendation performance for new users when there is insufficient data on new users, etc.)
• Reciprocal recommendation problem that satisfies both users
• Real-time recommendation problems that perform inference in a very short time on a set of recommendation candidates that change in real time (session-based recommendation, graph-based recommendation, reinforcment learning, …)
• Recommendation problem that considers trade-offs between multiple target indicators
• The problem of finding a primary target indicator that improves long-term indicators

[Trust & Safety]

We conduct research and development on various technologies to understand what content contains and utilize this information to provide a satisfactory user experience. We are looking for people who can solve the following problems together to extract useful information for decision-making by accepting unstructured data consisting of images, voices, and natural language as input. ( Squad Interview )

• Issues with lightweight models and optimizations that can achieve high speeds in mobile environments
• The problem of multi-task or multi-label models that are efficient and can control the importance of labels.
• Problems using partial multi-modal data
• The problem of detecting real-time abnormal users (e.g. spam/fake accounts) based on user behavior logs flowing into the stream and content understanding results.
• Efficient data labeling method through active learning or core-set selection method that can reduce the data required for model learning

[Generative AI]

Through various generative AI research and development, we provide users with unprecedented new experiences. We create tools that allow users to easily create personalized content and express themselves within the service, and develop new functions using generative AI. To this end, we are looking for people who can solve the following problems together. ( Squad Interview )

• Development of a personalized image generation model that can generate images of objects desired by users.
• Developing new features using large language models, learning, tuning and serving large language models for this purpose.
• Model development and optimization to ensure that large-scale generative models can reliably handle large amounts of traffic
• Research and consideration of how to innovate user experience within services by utilizing generative models

[common]

In common, we are also continuously working on researching AI technologies that are included in the product. In the actual production environment, there is no refined data set like Kaggle, and in most cases, new data is fed into the system every day. In order to build a Flywheel that automatically creates a better model today than yesterday, we are looking for people who can solve the following problems together.

• How to deal with highly imbalanced or noisy label data
• Continual/life-long learning method that can continuously improve existing deployed models
• A meta-learning method that can respond to changes in model task requirements and new services.
• Modeling, optimization, and distillation methods that can learn large-scale models and reliably process hundreds or thousands of inputs per second in a real service environment.

Eligibility

• Someone with a general understanding of the AI/ML domain and in-depth knowledge of at least one specific domain, and at least 3 years of related project experience
• Someone who can help build and maintain the team's technical competitiveness, including AI/ML, through continuous proactive learning.
• A person who has the ability to solve engineering constraints that cannot be solved by conventional methods, based on AI modeling capabilities and a deep level of understanding of software engineering in general.
• Anyone with experience integrating AI technology into real-world services and significantly improving key metrics
• Someone with strong communication skills and engineering capabilities who can collaborate with stakeholders across multiple disciplines to train ML models and deploy them to real services.
• Anyone with experience in quickly and accurately implementing a paper from scratch that has not yet been publicly disclosed
• Anyone with experience helping other ML-related engineers grow or has relevant capabilities
• Degree or nationality is not important, but those who can communicate fluently in Korean are welcome.

Preferential treatment

• Those who have published in top-tier machine learning conferences and journals (NeurIPS, ICLR, ICML, CVPR, ICCV/ECCV, KDD, etc.) or have won awards in AI-related competitions
• Anyone with experience with SotA on public benchmark data sets
• Someone with extensive development experience outside of AI/ML, including client (Android, iOS) and backend
• Anyone with experience participating in open source development related to machine learning
• Someone who can boast extensive knowledge across the AI/ML domain
• Someone who has experience in planning A/B test experiments, defining target KPI indicators, and conducting SQL-based data analysis
• Experience using causal analysis and other statistical techniques to extract meaningful insights from data and use them for decision making
• Someone with experience leading an engineering team
• Someone who is fluent in English