ML design, development, and productionization across various algorithms and problem types.

Managed complex enforcement projects and investigations for proactive finding of bad ads.

[Rotation] Built a semi-supervised learning for label propagation to propagate integrity labels from a seed set of labelled Google ads to the huge set of unlabeled Google ads based on the similarity.

[ML Bootcamp] Participated an intensive course about Machine Learning at Headquarters at Google, followed by a 3 month rotation in an ML engineer team. The program is designed to give participants the practical knowledge to apply ML to Google-scale problems. The Ninjas gain first-hand experience developing and launching ML models in Google.

Working on temporal database, temporal keyword search

Project participated: Effective Conceptual Relational Indexing. Application to Medical Domain with Multilingual Documents.- Project description: Concepts can be defined as human understandable unique abstract notions, independent from any languages or information representation. They are used to organize perception and knowledge. The project adds an extra step to document indexing by extracting concepts from documents. Application of the IR model and technical solution has to be done on a restricted domain: the medical domain, where knowledge set and ontology are already digitally available.- Technologies/ Resources used: CLEF, UMLS, Perl, X-IOTA