Dedicated graduated student in the field of computer science. Successful completion of advanced academic projects demonstrating engineering support capability. Aiming to improve and use my knowledge in data science, Big data, Backend engineering, computer vision, natural language processing in order to support senior managers and upper management in decision making for company.
Description: A General Game Player (GGP) designed engine which supports game of HEX right now. It uses reinforcement learning algorithm namely as Monte carlo tree search along with plenty of simulation strategies for improving search algorithm.
Description: An autonomous driver for the game of speed dreams. In this project desktop screenshots are appended to keyboard keys and collected as data. The data is fed to a reception v3-type CNN and after training, CNN learned how to keep a driving car between road lines
Description: The project splits in 2 parts. In first part initially we use data analysis and feature extraction for dataset. After that we define the best model that can use extracted features to predict the travel time with least possible error. In the second part we use genetic algorithm to find the path between random locations in test data which has the minimum distance.
Description: This is a simple project to implement a restful api which can Retrive/Add/Remove/Modify contacts. We using MVC design pattern and Test-Driven-Development methodology.
Description: An analysis of kaggle glass dataset as well as building a neural network. In this work Neural Network is built with considering optimized parameters using hyperopt and hyperas libraries. We also investigated the impact of using NearMiss and Imblearn methods for undersampling and Oversampling dataset.
Description: Neshan application is a unique iranian Intelligent Transportation System application which extract and estimate traffic and transportation data,and provides services like ITS and LBS to end-users. I was working with Rajman information structures to provide traffic data analysis services based on Urban ITS detectors/sensors.
Description: It's a step by step tutorial for students who wants to build models for NLP. It includes wrangling data, creating word2vec model, and implementing various models for comparison.
Description: In R&D process, this project was defined by senior manager to be used for both video surveillance and autonomous driving system. After researching through papers, github codes and other references, initial system were designed by using state-of-the-art Deepsort algorithm which is really fast and accurate.
Description: A basic model for video surveillance system which makes use of RetinaNet algorithm for finding face landmarks. The code could be run on both gpu and cpu which is a cool feature of Pytorch library.
Description: The role of lane detection system is really critical in autonomous driving system. 2 top algorithms for lane-line detection and segmentation are LaneNet and SCNN. we implemented initial version of both algorithms in our company.
Description: Enhancing the quality of simulations in Monte carlo tree search using quality-based rewards and considering simulation depth as a quality metric.
Description: In this article, I explained how I managed to build a module for logging the detections in real-time application. I managed to cover concepts of software development, unit-testing (TDD) and making use of that in Machine learning applications.