Data Scientist

The Internet is full of Scams, Spam, Malware, and Low-Quality pages. More and more of it is pumped out every minute by countless numbers of financially incentivized, highly creative, and resourceful bad actors. Do you want to put your creativity and ML, Analytical, and Engineering skills to good use by defending the World from such bad guys? Do you have the skillset, drive, determination, and _can-do_ attitude to significantly tip the balance in this very challenging adversarial space, where bad actors are always looking for ways past our protections and are quick to exploit any gaps?

The Index Quality group within WebXT at Microsoft is looking for a passionate Applied Scientist in core ML team. The team works on exciting ML and Engineering problems in NLP, graph algorithms, and Web Search. We use cutting-edge technologies in challenging areas of Document understanding using classical and Deep Learning, leverage the power of Large Language Models (LLMs); in the area of Web Graph, User graph, and Log analysis through Graph Embedding, cluster and community analysis, behavioral pattern mining; and in Web Ranking by modeling document quality, safety, authority. Our algorithms operate at Web Scale on hundreds of billions of pages, and at the same time analyze each fresh document as it is crawled. We utilize a combination of user behavior and feedback, sophisticated LLM agents, Crowdsourcing, and Expert raters to keep our defenses up-to-date and continually improve our models for the future.

Working in this team, you will have the opportunity to propose and build mechanisms that not only stop currently known bad practices from reaching our users, but also to set up proactive protections against future, not-yet-seen, novel attacks; all while ensuring that good pages are unaffected by filters. You will be part of a team that is responsible to provide a clean, safe, high quality, and comprehensive subset of the Internet to more than a billion users through various Microsoft offering such as Bing, Microsoft News and Feed, Edge, Windows, and Office 365; as well as external partners relying on our APIs and Search solution.

Why work at WebXT

Inside Microsoft's Web Experiences Team

**Responsibilities**:

- Innovating, designing, implementation, execution, and maintenance of NLP/ML models and algorithms focusing on safety and quality.
- Build feature engineering pipelines, design and execute training data lifecycle, train+test+deploy classical as well as machine learning/deep learning models and heuristics.
- Analyze and keep updated with trends of black-hat SEO and spammy behaviors, and malware.
- Build techniques to detect sophisticated spam behaviors using Web Data Platform infrastructure in combination with expertly-tuned Large Language Models (LLM).

**Qualifications**:

- ** Basic Qualifications**:

- 1+ year experience in machine learning, deep learning and/or related fields.
- BS/MS in Computer Science, Statistics, Applied Mathematics, Physics, or other engineering or science fields and 2+ years industry experience in related fields OR PhD with 1+ years industry experience.
- Experience with C#/C++/Java/Python with a good practical grasp of Data Structures and Algorithms.

**Preferred Qualifications**:

- Graduate degree with experience in Machine Learning.
- Experience developing end to end ML/DL systems.
- Have research or work experience on NLP, solving complex problems through LLM prompt engineering, Inferences on Graphs, recommender system.
- Familiarity with how Search Engines work and common SEO techniques
- Proficient in building with and deploying models in deep learning computational graph frameworks such as Tensorflow/Pytorch/MXNet.
- Have publications at peer-reviewed AI conferences (e.g. NIPS, CVPR, ICML, ICLR, ICCV, and ACL).

Backend# #Search# #Safety# #MachineLearning# #LLM# #NLP#