Mahsa Lotfi

Email: adjlx0@r.postjobfree.com GitHub Repository: https://github.com/Mahsalo

LinkedIn: http://www.linkedin.com/in/mahsa-lotfi Personal Webpage: https://mahsalo.github.io/ Phone: +1-972-***-****

Stanford University, United States (2019-Present)

Post Doctoral Researcher- Statistics Advisor: David Donoho The University of Texas at Dallas, United States (2015- 2018, GPA: 3.95) PhD in Electrical Engineering - Signal Processing Advisor: Mathukumalli Vidyasagar Isfahan University of Technology, Iran

M.Sc. in Electrical Engineering - Signal Processing (2012-2015, GPA: 3.6) B.Sc. in Electrical Engineering, Signal Processing (2008-2012, GPA: 3.5) Super-Resolution Fluorescence Microscopy in BEAM Imaging

- Proposed and implemented a multiple-frame compressed sensing recovery approach in resolving the low-resolution images in BEAM imaging.

- Implementations are in Python and MATLAB and the models are analyzed in R.

- Parallel computing done using ClusterJob and ElastiCluster frameworks for HPC cluster at Stanford, Google Cloud and AWS virtual machine instances. Hands-On Unsupervised Learning with TensorFlow 2.0

- Conducted a comprehensive hands-on video course on unsupervised learning including Clustering (K-Means, Hierarchical, t-SNE, DBSCAN), Principal Component Analysis

(PCA), Anomaly Detection, Autoencoders, Deep Belief Networks, Generative Adversarial Networks (GANs) and Self Organizing Maps.

Classification, Feature Extraction & Automatic Detection of Iron Deficiency Anemia

- Designed a machine learning approach including segmentation, feature extraction and classification of 10 different cell types of RBCs using SVM, KNN and decision trees.

- Applied a 4-layer Deep Convolutional Neural Network with Drop-out and Max-Pooling layers to detect the cell-types and tackle overfitting. (MATLAB and Python/Keras) Sparse Regression Application in Ovarian Cancer

- Proposed a novel approach in the prediction of time to tumor recurrence in Ovarian cancer using sparse regression.

- Applied Combined-L1-and-L2 (CLOT), LASSO and Elastic Net and increased the robustness in the sparse recovery while keeping the number of final features small by using CLOT. (MATLAB)

Design of a Recommender System Using PySpark on Data Bricks

- Designed a recommender system based on customer’s previous behaviors for RetailRocket dataset from Kaggle. Used ALS for Matrix Factorization. Signal Recovery and Data Compression in Compressive Sensing

- Proposed an innovative approach in compressive sensing using binary sensing matrices based on Array LDPC codes parity check matrices and Euler Square matrices. (MATLAB)

- Optimized the CPU time and storage requirement in the recovery of high-dimensional sparse signals.

- Designed a super-fast, non-iterative signal recovery algorithm based on binary sensing matrices which is hundreds of times faster than basis pursuit algorithm. (MATLAB) Python

TensorFlow

Keras

MATLAB

R

SQL

Tableau

ElastiCluster

Google Cloud

ClusterJob

Parallel Computing

Unix/Linux

C++

EDUCATION:

PROJECTS:

SKILLS:

From Data to Insights with Google Cloud Platform (Specialization), Coursera, 2020 Fundamentals of Visualization with Tableau, Coursera, 2020 R Programming, Coursera, 2020

SQL for Data Science, Coursera, 2020

Google Cloud Platform Big Data and ML Fundamentals, Coursera, 2020 ICME Workshop on Data Science (Deep Learning, Optimization, NLP, Statistics), 2019 Deep Learning with TensorFlow, University of Texas at Dallas, 2018 Big Data, Big Data Club at University of Texas at Dallas, 2017 Mahsa Lotfi and M. Vidyasagar, “Compressed Sensing Using Binary Matrices of Nearly Optimal Dimensions”, IEEE Transactions on Signal Processing, April 2020. Mahsa Lotfi and M. Vidyasagar, “A Fast Non-iterative Algorithm for Compressive Sensing Using Binary Measurement Matrices”, IEEE Transactions on Signal Processing, 66(15), pp. 4079-4089, August 2019. Mahsa Lotfi, Burook Misganaw and M. Vidyasagar, “Prediction of Time to Tumor Recurrence in Ovarian Cancer: Comparison of Three Sparse Regression Methods”, Bioinformatics Research and Applications, 1-11, 2017 Mahsa Lotfi and M. Vidyasagar, “Compressed Sensing with Binary Matrices: New Bounds on the Number of Measurements”, 5th Indian Control Conference (ICC), India, pp.17-21, January 2019. M. Vidyasagar and Mahsa Lotfi, “Compressive Sensing and Algebraic Coding: Connections and Challenges”, Uncertainty in Complex Networked Systems, pp. 275-322, 2018. Mahsa Lotfi and M. Vidyasagar, “Array LDPC Code-based Compressive Sensing”, 56th Annual Allerton Conference on Communication, Control and Computing, IL, USA, pp. 682-685, October 2018. Mahsa Lotfi and M. Vidyasagar, “Exact Recovery of Sparse Signals Using Binary Measurements”, Indian Control Conference, Kanpur, India, pp. 83-88, January 2018. Mahsa Lotfi and M. Vidyasagar, “A Fast Single-Pass Algorithm for Compressive Sensing Based on Binary Measurement Matrices”, 55th Annual Allerton Conference on Communication, Control and Computing, IL, USA, pp. 369-373, October 2017.

Mahsa Lotfi, B. Nazari, S. Sadri and N. Karimian Sichani, “The Detection of Dacrocyte, Schistocyte and Elliptocyte cells in Iron Deficiency Anemia, 2nd International Conference on Pattern Recognition and Image Analysis, Guilan, Iran, March 2015.

-Jan Van der Ziel Graduate Fellowship- UT Dallas, Fall 2018

-Louis J. Beecherl Graduate Fellowship- UT Dallas, Fall 2017 David L. Donoho Mathukumalli Vidyasagar Masaaki Nagahara Stanford University University of Texas at Dallas Kitakyushu University

+1-650-***-**** Indian Inst. of Tech. Hyderabad adjlx0@r.postjobfree.com adjlx0@r.postjobfree.com adjlx0@r.postjobfree.com

CERTIFICATES:

PUBLICATIONS:

HONORS/AWARDS:

REFERENCES:

Contact this candidate