Background & experience
10 years of professional software development and optimization experience.
Software engineer focused on designing and building reliable, high-performance systems. I enjoy owning problems end-to-end, from architecture and implementation through performance optimization and production hardening. I’m strongest when working on complex technical challenges that require both deep systems knowledge and clear collaboration across teams.
Azure Core · Microsoft
Led architecture and delivery of a diagnostics platform for a 1000× scale-up of Azure Storage deployments, while optimizing health/quorum systems (9× lower resource use, millions of targets). Also delivered disk management C++ libraries, live network isolation, redesigned auth, and a Python test framework that cut sign-off from hours to minutes.
Paycom
Built and shipped tax-reporting features in C#/.NET. Optimized large-dataset processing by 35%, created a reflection-based library that cut new-form development time by 40%, and reduced client tickets 20% through root-cause fixes.
Learn Lab · The University of Texas at Arlington
Built a hierarchical path planner that used a TensorFlow CNN to generate heuristics for A*/RRT, cutting planning time 60% on large maps. Accelerated the training/testing pipeline 80% via multi-processing and automated hyperparameter search.
XPeng Motors
Developed a C++/Python perception pipeline that improved performance 50% and supported multi-environment deployment. Optimized AI model inference with the ML team and built debugging/monitoring tools for the module and its communications.
Robotics Lab · Ferdowsi University of Mashhad
Created a robust omnidirectional walk engine for NAO robots (refined with evolutionary algorithms) and an Artificial Potential Field planner for real-time navigation. Extended the RoboCup 2D simulator for multi-agent research scenarios.
Menulit Inc.
Designed and built a full-stack restaurant menu and ordering system (C#/EF/MySQL, Android app, AngularJS portal). Focused on reliability through edge-case testing and aligned features with real restaurant workflows.
Ferdowsi University of Mashhad
Courses: Artificial Intelligence, Operating Systems, Compiler Design Concepts
Ferdowsi University of Mashhad
Course: Introduction to Linux, C++, and RoboCup 2D soccer simulation
The University of Texas at Arlington
Thesis: Learning Hierarchical Traversability Representation for Efficient Multi-Resolution Path Planning. Supervisor: Prof. Manfred Huber.
Ferdowsi University of Mashhad
This work introduces a learning framework using a deep neural network with a stackable convolution kernel to establish a hierarchy of directional traversability representations with decreasing resolution that can serve as an efficient heuristic to guide a multi-resolution path planner.
Developed an RGB video-based system for recognizing gym exercises. Extracted features from video and classified activities using SVM, Decision Tree, KNN, and Random Forest, with Random Forest achieving 98.98% accuracy. Added repetition counting via local minima and form evaluation with Dynamic Time Warping, plus an interactive Android app for workout insights.