Daria Orhunova
Engineering Manager / Data Scientist / Machine Learning Engineer at Adimen- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
Topline Score
Bio
Experience
-
Adimen
-
Ukraine
-
IT Services and IT Consulting
-
1 - 100 Employee
-
Engineering Manager / Data Scientist / Machine Learning Engineer
-
Sep 2020 - Present
- Implemented a multilingual intent classification pipeline for chatbots applications. - Created a multilingual text clustering tool for self-supervised labelling; using this tool we partially automated the taxonomy generation task for raw texts. - Set up data storage and data versioning processes along with versioning and managing experiments. Stack: NumPy, pandas, scikit-learn, PyTorch, transformers, spacy, NLTK, faiss, bash, Docker - Implemented a multilingual intent classification pipeline for chatbots applications. - Created a multilingual text clustering tool for self-supervised labelling; using this tool we partially automated the taxonomy generation task for raw texts. - Set up data storage and data versioning processes along with versioning and managing experiments. Stack: NumPy, pandas, scikit-learn, PyTorch, transformers, spacy, NLTK, faiss, bash, Docker
-
-
-
DataMix
-
Ukraine
-
IT Services and IT Consulting
-
1 - 100 Employee
-
Lead Data Scientist
-
Mar 2020 - Aug 2020
- Developed a news clustering model using a Transformer-like model. - Implemented several models for licence plate recognition. Stack: NumPy, pandas, scikit-learn, PyTorch, transformers, spacy, NLTK, mmdetection, Docker, flask - Developed a news clustering model using a Transformer-like model. - Implemented several models for licence plate recognition. Stack: NumPy, pandas, scikit-learn, PyTorch, transformers, spacy, NLTK, mmdetection, Docker, flask
-
-
-
Samsung
-
Киев, Украина
-
Machine Learning Engineer (NLP)
-
Jul 2019 - Mar 2020
- Developed a data augmentation pipeline for NER models for cybersecurity domain. - Refactored a data preparation tool, resulting in a lower number of corrupted texts which increased a NER model performance by 3-5% depending on the entity type. - Implemented an event detection pipeline for twitter data. Stack: NumPy, pandas, scikit-learn, PyTorch, flair, transformers, spacy, NLTK - Developed a data augmentation pipeline for NER models for cybersecurity domain. - Refactored a data preparation tool, resulting in a lower number of corrupted texts which increased a NER model performance by 3-5% depending on the entity type. - Implemented an event detection pipeline for twitter data. Stack: NumPy, pandas, scikit-learn, PyTorch, flair, transformers, spacy, NLTK
-
-
-
LinkAlong
-
Switzerland
-
IT Services and IT Consulting
-
1 - 100 Employee
-
Data scientist (NLP)
-
Jun 2018 - Jun 2019
- Prepared several datasets for evaluation of text clusterization approaches. - Extended the existing topic detection system with 3 new languages (Chinese, Arabic, Spanish). - Adapted a tensorflow-based BERT model for multilabel topic classification, outperforming the previously existing rule-based approach by a significant margin. Stack: NumPy, pandas, scikit-learn, NLTK, Stanford NLP, spacy, TensorFlow, Docker - Prepared several datasets for evaluation of text clusterization approaches. - Extended the existing topic detection system with 3 new languages (Chinese, Arabic, Spanish). - Adapted a tensorflow-based BERT model for multilabel topic classification, outperforming the previously existing rule-based approach by a significant margin. Stack: NumPy, pandas, scikit-learn, NLTK, Stanford NLP, spacy, TensorFlow, Docker
-
-
-
DataRoot Labs
-
Ukraine
-
Software Development
-
1 - 100 Employee
-
Data science Trainee
-
Sep 2017 - Mar 2018
- As a final project, implemented a UNet-based model for retina vessels segmentation. Stack: NumPy, pandas, scikit-learn, TensorFlow, flask - As a final project, implemented a UNet-based model for retina vessels segmentation. Stack: NumPy, pandas, scikit-learn, TensorFlow, flask
-
-
-
Materialise
-
Belgium
-
Software Development
-
700 & Above Employee
-
Medical Engineer
-
Mar 2017 - Jan 2018
- Created several hundreds of personalized 3D models of human craniums based on CT scans. - Prepared a number of plannings for maxillofacial surgeries. - Created several hundreds of personalized 3D models of human craniums based on CT scans. - Prepared a number of plannings for maxillofacial surgeries.
-
-
Education
-
National Technical University of Ukraine 'Kyiv Polytechnic Institute'
Master's degree, Applied physics and Nanotechnology -
National Technical University of Ukraine 'Kyiv Polytechnic Institute'
Bachelor's degree, Applied physics