Protect TikTok users, including and beyond content consumers, creators, advertisers. Secure platform health and authenticity of community experience. Collaborate with cross-functional stakeholders to improve TikTok infrastructures, services, tools and algorithms, towards a higher standard of privacy and security.
Requirements
- Solid engineering skills. Proficiency in at least two of: Linux, Hadoop, Hive, Spark, Storm.
- Strong machine learning background. Proficiency or publications in modern machine learning theories and applications such as deep neural nets, transfer/multi-task learning, reinforcement learning, time series or graph unsupervised learning.
- Research experience in one or more of the following fields: applied machine learning, machine learning infrastructure, large-scale recommendation system, market-facing machine learning product.
- Strong publications record in top conferences or journals (e.g., NeurIPS, ICML, ICLR, CVPR, etc) and journals (e.g., TPAMI, JMLR)
- Track record of high impact research.
Responsibilities
- Build machine learning solutions to respond to and mitigate business risks in ByteDance products/platforms. Such risks include and are not limited to abusive accounts, fake engagements, spammy redirection, scraping, fraud, etc.
- Improve modeling infrastructures, labels, features and algorithms towards robustness, automation and generalization, reduce modeling and operational load on risk adversaries and new product/risk ramping-ups.
- Level up risk machine learning excellence on privacy/compliance, interpretability, risk perception and analysis.
Other
- Currently pursuing a Bachelor or above degree in computer science, statistics, or other relevant, machine-learning-heavy majors.
- Ability to think critically, objectively, rationally. Reason and communicate in result-oriented, data-driven manner. High autonomy.
- Internships at TikTok aim to offer students industry exposure and hands-on experience.
- It runs for 12 weeks.
- Please state your availability clearly in your resume (Start date, End date).
