Ali Azam
****.***.****@*****.*** 605-***-**** linkedin.com/in/mali17/ github.com/aali17
SOFTWARE ENGINEER
Extensive experience in image processing and computer vision with OpenCV, Skimage, etc
In-depth knowledge in computer vision sensors (LIDAR, Optical Camera)
Hands-on experience in deep learning with tensorflow, keras, CUDA in Nvidia GPU
Experience with applied machine learning techniques e.g., statistical pattern recognition, principle component analysis (PCA), regression, k-means clustering, etc
Worked with deep learning techniques e.g., convolutional neural networks (CNN)
Hands-on experience in feature detection & matching, object tracking, 3D reconstruction, image segmenta- tion, classification, etc
Strong mathematical skills, including linear algebra, numerical methods, probability, and stochastic process
Substantial knowledge in optimization techniques e.g., gradient descent, simulated annealing, particle swarm optimization, genetic algorithm, etc
Familiar with core problems in robotics including state estimation (Kalman filter, particle filter), SLAM, etc
Research experience in autonomous systems, decision-making, computer vision, and sensor fusion SKILLS
Languages: Python, C++, and SQL
Algorithm development environments: MATLAB, Simulink, and Mathematica Libraries: Numpy, Matplotlib, TensorFlow, Keras, SciPy, Skimage, Scikit-learn, OpenCV, PIL, and CUDA Software tools: Spyder, Jupyter, PSpice, Anaconda, CodeBlocks, Arduino, and Linux (Ubuntu) Design automation tools: PSpice, Multisim, and Proteus Network simulation: Packet Tracer
Test and measurements: signal generators, oscilloscopes, spectrum analyzer, and DMM Others: HTML, CSS, and LATE
X
EXPERIENCE
South Dakota School of Mines and Technology Rapid City, SD Graduate Research Assistant Aug. 2018 - Dec. 2020
UAV control optimization for multi-agent swarm systems for applications e.g., formation control, and target tracking
Posed decentralized Markov decision processes (Dec-MDP) for UAV swarm guidance and control
UAV swarm control for 3D reconstruction of a scene using multiview stereo (MVS) and structure from motion (SfM) Summer Robotics Camp Assistant May 2019 - June 2019
Assisted students with C++ code to program and upload to Arduino UNO board to control Elegoo 3.0 car
Prepared lessons, helped students with programming class, arranged programming contest, and evaluated students
Graduate Teaching Assistant Aug. 2018 - Dec. 2018
Helped students in the lab to make sure the experiment is being conducted properly
Graded homework, lab report, and maintained office hours for questions from students related to the course Fair and Appropriate Technology Limited Gazipur, Bangladesh System Engineer Mar. 2016 - Nov. 2016
Configured 2G, 3G, and microwave links, and investigated problems, diagnosed and repaired faults of RBS
Conducted drive test (DT) and preliminary acceptance test (PAT)
Supervised engineering and technical staff, and scheduled and coordinated work to tight deadlines
Maintained fiber links and all types of operational and maintenance infrastructure in the region PROJECTS
Traffic Light Classification link to project
Used classical computer vision techniques feature extraction and description to classify traffic lights (accuracy: 90.23%)
Used community standard convolutional neural network (CNN) for traffic light classification (accuracy: 98.99%)
Feature based approach could make sure no red traffic signal is classified as green, CNN failed to do so Statistical Pattern Recognition link to project
Computed and analyzed the eigenspace of given image datasets and appropriate subspace dimension
Performed object classification in a dataset of 20 object class and pose estimation in 128 different poses of each object
(pose estimation accuracy: 99.64%)
Feature Detection and Matching link to project
Extracted features and edges, and corresponded said features between successive image frames using Harris feature detector, Laplacian of Gaussian detector, and SIFT
The detected features and developed descriptors were invariant to rotation, scale, and illumination changes Image Filtering and Hybrid Imaging link to project
Frequency domain: blended high and low frequency content of two images to create a hybrid image that was perceived differently at different distances
Spatial domain: created my own convolution function to apply low and high frequency Gaussian kernel for image filtering and blended them to create a hybrid image EDUCATION
South Dakota School of Mines and Technology Rapid City, SD MS in EE with minor in CS Dec. 2020
Rajshahi University of Engineering and Technology Rajshahi, Bangladesh BSc in Electronics and Telecommunication Engineering Dec. 2014 ONLINE LEARNING
Intro to Self-Driving Cars Udacity Apr. 2020
Learned object oriented programming, linear algebra, translated Python program to C++, and optimizing codes
Learned frequently used data structures and algorithms, popular visualization libraries of Python, computer vision, and machine learning
Self-Driving Car Engineer Udacity Ongoing
Learning computer vision and deep learning to automotive problems e.g., detecting lane lines, predicting steering angles
Learning sensor fusion to filter data from an array of sensors in order to perceive the environment Convolutional Neural Networks Coursera Feb. 2020
Learned how to build a convolutional neural network (CNN) and applied CNN to visual detection and recognition tasks
Learned how to apply CNNs to a variety of image, video, and other 2D or 3D data COURSEWORK
Graduate
Introduction to Computer Vision, Machine Learning Fundamentals, Linear System Theory, Optimization Techniques, Advanced Topics in Artificial Intelligence: Natural Computing, Robotic Control System, and Intelligent Control System Undergraduate
Computer Fundamentals and Programming, Data Structure and Algorithms, Digital Image Processing, and Digital Signal Processing