Data Science

Programming Languages: Python, SQL, NoSQL, MATLAB/Octave, C++, R (some experience)

Tools & Frameworks: NumPy, SciPy, Matplotlib, Scikit-Learn, Deep Learning, NLP, Computer Vision, Git, Sequence Models, OriginPro, OpenCV, TensorFlow, Keras, UNIX, Tableau, spaCy, kNN, SVM, statsmodels Data Science Fellow,Insight Data Science, Silicon Valley, CA Sep 2020 - Present

● Designed a smart filter for social media influencers to automatically replace company logos and branded symbols in their social media posts

● Detected company logos in images and videos using YOLOv5 trained on QMUL-OpenLogo dataset (0.6 mAP@0.5), replaced logos using generative image inpainting, and auto-posted on social media account

● Built a web extension for influencers to implement this filter automatically (GitHub) Research Data Scientist, Sandler Neurosciences Center, UCSF Jan 2017 - Aug 2020

● Designed experiments to understand the structure and function of neural circuits

● Collected, cleaned, and validated data to ensure accuracy and consistency

● Implemented a first-of-its-kind study that embedded neural data using diffusion maps to compare against Alexnet, VGG-16, and Inception-3 networks, resulting in new findings about deep convolutional neural networks (CNNs)

● Mentored two postdoc and graduate students in data acquisition & analysis, resulting in 10+ publications

● Served as President of the Iranian Neuroscientist Community for two years; secured funds for travel awards benefiting international researchers and organized networking events Random Name Generator Using NLP and SpaCy I UCSF,GitHub Developed a character-level language model in tensorflow to generate new names COVID-19 CT Segmentation GitHub

Developed a clinically-applicable Python package to implement CNNs for accurate diagnosis and segmentation of lungs and pneumonia-like infections (0.86 dice coeff) using CT images of COVID-19 patients Decoding Stimulus Identity GitHub

Developed neural networks in order to decode stimulus identity (direction of motion, size, etc.) in neuronal recordings from mouse visual cortex (mean accuracy>0.4, chance level 0.16) (bioRxiv) Ph.D. in Physics, Washington University in Saint Louis Oct 2016 M.Sc. in Physics, Sharif University of Technology May 2010 Mahmood Hoseini

San Francisco, CA +1-314-***-****

*******.********@*****.***

linkedin.com/in/Mahmood-Hoseini

github.com/Mahmood-Hoseini

SKILLS

EXPERIENCE

PROJECTS

EDUCATION

Contact this candidate