Marija Stanojevic, Phd Email & Phone Number

Canada

Toronto, Ontario, Canada

Leading machine learning team and projects. Technical design of new products. Fundraising (LLMs, real-time applications, team project lead, system design).

Toronto, Ontario, Canada

- Led the development of a multimodal, multilingual automatic framework designed to predict the severity score and category of a disease from a speech sample in an open-ended task. (Transformers, CNN, PyTorch, Docker, AWS, Research).- Led machine learning research initiative including organization of Machine Learning for Cognitive and Mental Health.

Philadelphia, Pennsylvania, United States

https://scholar.google.com/citations?user=pAyfhIkAAAAJ&hl=en- RA (Sep-May 2017/18, 2020/21): NSF, NIH, CDC, and IQVIA funded projects.- TA (Sep-May 2018/19, 2020/21, 2021/22). Courses: 1) Knowledge Discovery and Data Mining, 2) Computer and Low-Level Programming, and 3) Data Structures.- Presidential Fellow (Jan 2017 - Aug 2020): awarded based on previous.

Sunnyvale, California, United States

- Proposed and implemented neural collaborative filtering with multimodal longitudinal representation learning for user embedding (Spark, Scala, Keras, Tensorflow, Dali).

New York, New York, United States

- Designed and implemented a novel multitask multilabel multimodal attention architecture for extreme classification (Python, Presto, PyTorch, Caffe2, DL, Transformers, internal tools).

Menlo Park, California, United States

- Recruiting Science: Improved candidate search by implementing NLP and IR techniques to reduce long tail in skills distribution and by proposing, implementing, and evaluating a novel Siamese-like architecture to embed job descriptions (Python, Presto, PyTorch, Caffe2, DL, statistical NLP, IR, internal tools).

Chicago, Illinois, United States

- Worked at Epsilon, Conversant division, Decision Sciences Team- Pioneered a solution to a large-volume spatio-temporal problem utilizing mean, quick-shift, and hdbscan clustering. Created a proxy to test existing product.- Defined evaluation metrics to show potential for implementation into a product (Hadoop, Hive, python, pandas, geo, folium, geopandas.

Belgrade, Serbia

- Led school performance insight software development, achieving a 30\% improvement in efficiency through optimized data cleaning and integration.- Reduced costs by 50\% by implementing in-house data science and data visualization techniques (PHP, ETL, AWS, R, Python, MySQL, D3.js).

Doctor Of Philosophy (Phd), Machine Learning, Gpa: 3.94/4.0

Courses: - Complex networks - Principles of Data Management - Machine Learning - Computer Vision - Statistical Methods I and II.

Master'S Degree, Signal Processing With Focus On Machine Learning And Pattern Recognition

Courses: - Statistical Classification of Signals - Pattern Recognition - Speech Recognition - Soft-computing - Robotic Systems Thesis.

Bachelor Of Engineering (Be), Computer Software Engineering

Courses: - Algorithms and Data structure (I and II with lab projects, C++) - Math (I and II) and applied math (I and II) courses.