[Publishing Platform Div.] Game Security Data Scientist (7년 이상)
About this role
Data scientists who enjoy untangling complex behavioral patterns and building robust detection frameworks will find a natural home within this group. Operating at KRAFTON, this security-focused organization develops standard platforms and operational systems to safeguard global gaming environments against unfair play and abnormal user behavior.
This practitioner will focus on advancing machine learning capabilities within the KRAFTON Security Services division by designing and refining detection models for anomalies and cheating activities. Daily responsibilities involve processing large volumes of game logs, engineering relevant features, and ensuring detection results directly support security operations and sanction workflows.
Ideal candidates bring substantial background in statistical modeling, possess strong data wrangling abilities using Python or SQL, and thrive on defining their own problems to drive data-led improvements. Success in this role requires close collaboration with game studios and security engineers to evaluate model accuracy and reduce false positives.
What you will do
- Develop and improve machine learning models to detect abnormal user behavior and cheating within games.
- Analyze massive game logs, user behavior data, and security events to surface early indicators of anomalies.
- Analyze cheating patterns, extract detection features, and design advanced detection logic.
- Provide dedicated data analysis support for ongoing security operations and sanction response procedures.
- Design and visualize operational metrics to increase the overall efficiency of detection systems.
- Collaborate directly with game studios and security engineers to align technical solutions with operational needs.
- Evaluate model performance quantitatively while analyzing and mitigating false positive and false negative results.
What you need
- Possess a degree in Machine Learning, Deep Learning, Statistics, Computer Science, or a related field, or have equivalent practical experience.
- Demonstrate hands-on experience with machine learning modeling projects covering anomaly detection, classification, prediction, or pattern analysis.
- Utilize Python and SQL proficiently to analyze large-scale log and event data.
- Understand the complete data pipeline from log data refinement and feature engineering to dataset construction and model evaluation.
- Comprehend machine learning and deep learning model architectures well enough to improve them according to specific operational goals.
- Apply solid feature engineering capabilities that account for data quality, reproducibility, and scalability.
- Quantitatively evaluate model performance and derive improvement directions from false positive and false negative perspectives.
- Demonstrate the ability to independently define problems and drive improvements based on data.
Nice to have
- Experience applying game data toward anti-cheat, cheating or abuzing detection, and anomaly analysis
- Firsthand play experience or understanding of data characteristics across game genres such as FPS/TPS, battle royale, and MMORPG
- Experience detecting abnormal patterns based on user behavior data and improving models or detection logic
- Experience processing large-scale log data and building or operating machine learning pipelines
- Familiarity with large-scale data platforms like Spark and Hadoop
- Experience collaborating with global organizations and studios or communication skills in English or a second language
Skills and tech stack
Questions candidates ask
What is the work location for this position?
The role is based at the Yeoksam Centerfield office in Seoul.
What does the hiring process look like?
The evaluation consists of a document screening, phone interview, take-home test, personality assessment, technical fit interview, and 2 to 3 rounds of culture fit interviews.
What are the employment conditions regarding probation?
A 5-month probationary period applies during which employment may be terminated early or not continued based on evaluation results, with no changes in employment type or salary during this time.