Responsibilities
[About ByteDance]
Founded in 2012, ByteDance's mission is to inspire creativity and enrich life. With a suite of more than a dozen products, including TikTok, Helo, and Resso, as well as platforms specific to the China market, including Toutiao, Douyin, and Xigua, ByteDance has made it easier and more fun for people to connect with, consume, and create content.
[About the Team]
Our Trust and Safety Engineering team is responsible for building a highly scalable, complex and intelligent platform that is able to manage high traffic loads. You would also be building the tools to help protect our users from harmful content and abusive behaviors. With the continuous efforts of our trust and safety team, we can provide the best user experience and bring joy to everyone in the world.
[Responsibilities]
- Conduct offline and online experiments, analyze according data and evaluate to improve algorithms and strategies;
- Responsible for the research and development of cutting-edge machine learning technology, application of visual, speech, audio and text recognition post-processing technology to products
- Responsible for the optimization of NLP/Computer Vision/MultiModal algorithms for content understanding, and applying them to the scenarios of Trust and Safety.
- Responsible for building the ML model to recognize inappropriate content in the Trust and Safety context
Qualifications
[Minimum Qualifications]
- Bachelor's degree in Computer Science or equivalent with 3+ years of experience
- Familiar with large-scale language modeling algorithms (N-gram, nnlm)
- Familiar with decoding engine optimization, with experience in WFST composition and optimization
- Familiar with decoding algorithms under the end-to-end framework, with experience in decoding algorithm development and optimization
- Solid experience with data structures and algorithms;
- Software development experience through hands on coding in a general purpose programming language;
- Experience in one or more of the following areas: Machine Learning, Recommendation Systems, Data Mining or other related areas;
- Familiar with one or more of the open source tools such as Kaldi, SRILM, RNNLM, Pytorch, etc.
- Experience in acoustic models, language models, speech recognition post-processing (punctuation prediction, inverse text normalization, disfluency detection, error correction), natural language processing or other related algorithms.
- Possess Linux environment R&D capabilities, proficient in Python/Go/C++.
- Prior experience working with Internet Products or large scale platforms.
- Prior experience in Content Moderation or Trust and Safety Function
[Preferred Qualifications]
- Fast business understanding and collaborative in teamwork;
- Publications in international conferences or journals (ICASSP, Interspeech, ASRU) is a plus
- Excellent analysis and problem-solving skills, passion for challenging problems.
- Have a good team spirit and strong communication skills.
- Proactive and enthusiastic, with demonstrated ability to work in a collaborative environment
