Mobina Jamali

I'm a roboticist and AI researcher working on Multi-Agent Reinforcement Learning.

I received my BSc in Physics from the University of Calgary in 2023, focusing on developing advanced architectures for automation and control of Autonomous Underwater Vehicles, advised by Alex Ramirez Serrano. After graduation, I started my self learning journey in the ever growing world of Artificial Intelligence. I completed a MicroMasters program in Data Science from the University of California SanDiego, learning about mathematical and applied sides of data science, while advancing my robotics knowledge by taking courses in The Construct platform .

Since then, I've engaged in independent and collaborative research projects with different universities, aiming to develop intelligent agents capable of supporting humans in executing complex, multi-step tasks through a combination of learning and planning. Currently, I’m advancing this work at Intelligent Robot Learning Lab under the supervision of Matthew E. Taylor, where I focus on human-in-the-loop multi-agent reinforcement learning.

Email / CV / linkedIn / Github

Teaching

[Winter 2022] Teaching Assistant, PHYS229 (Modern Physics), University of Calgary
[Winter 2022] Teaching Assistant, PHYS259 (Electricity and Magnetism), University of Calgary

Awards

[Spring 2022] Program for Undergraduate Research Experience (PURE), University of Calgary
[Spring 2020] BSc Summer Research Funding, University of Calgary
[Fall 2019] President's Admission Scholarship, University of Calgary

Research

I'm interested in embodied AI, deep learning, and computer vision. My goal is to enhance human-agent interaction in dynamic and unpredictable environments.

	Advancing the Control for a Highly Maneuverable Autonomous Underwater Robot Code / Thesis This project, completed for my undergraduate thesis at the University of Calgary, presents a comprehensive architecture for the dynamic modeling, control, and simulation of a Highly Maneuverable Autonomous Underwater Vehicle (HM-AUV).
	Robotic Arm Manipulation with Hierarchical Reinforcement Learning and Imitation Learning Code Solving multi-stage, long-horizon robotic manipulation tasks via Imitation and Hierarchical Reinforcement Learning.
	Super Mario AI Double Deep Q-Network Code Building Double Deep Q Network (DDQN) from scratch and training AI agent to Play Super Mario Bros.
	Document Scanner Code Creating real-time document scanner with great precision using computer vision.
	Cosmic Explorer Code Exploring the cosmos through analyzing the SDSS large-scale astronomical survey and training classifiers to explore celestial objects based on their properties.
	Chat-Mingle Telegram Bot Code Building a Telegram Bot capable of connecting random individuals together from diverse backgrounds giving them the possibility to chat.

Hobbies

	Hiking Living in beautiful Alberta means the Canadian Rockies are practically in my backyard. And take it from me- few things compare to the simple joy of enjoying a cup of tea at the summit after a challenging climb.
	Photography My lens helps me capture a piece of nature's beauty, from the fine details of a leaf to the vast, changing views of mountain ranges.

Fun Fact!

My inner curiosity at the age of 16 made me to spend two years on a project during high school, "Investigating the Effect of Arbuscular Mycorrhizal Fungi (Glomus etonicatum) and Air Pollutants on Growth Parameters of Maize (Zea mays L.)". Published in the Journal of Clinical Epigenetics, it led to several speaking invitations and presenting my work at Applied Microbiology and Beneficial Microbes in Osaka, Japan .
While this research experience was vastly different from my current work, but it was a huge jump for me to believe in what I'm doing and to always have the courage to pursue the unknown!

Design and source code from Jon Barron's website.