During my internship at KUIS AI LAB, I worked as the primary author on developing a novel framework for modeling videos using a pre-trained image generator's and a learned latent ODE, enabling a range of applications with high-resolution videos such as image animation, guided motion and appearance manipulation, and temporal interpolation and extrapolation. This work was accepted to ICCV 2023. For more details, see the project page
For my graduation project, I collaborated with my peers to develop a video chatbot Android application that incorporates cutting-edge AI models to offer a tailored and interactive language-learning experience. I am now volunteering with the Linguistics Justice League to scale our application and deliver it to refugee and immigrant youth living in the United States.
In this ongoing research project at KUIS AI LAB under the supervision of Prof. Aykut Erdem, I am investigating how language can be used to learn an image representation that will facilitate more natural image editing.The project is in its final stage and planned to be submitted in March 2023
During my research internship at the IUI LAB, I worked under the supervision of Prof. Metin Sezgin on designing a self-supervised feedback system for sketch-based open-ended questions. The project is based on Sketchformer for object-to-object alignment and a clustering-based approach for object-level segmentation. The shared video is a sample of transforming student answer (bottom-left corner) to the correct answer (upper-right corner)
Opera is an offline simulator for distributed systems. Opera provides an integrative and easy-to-use interface for running, testing, and analysing distributed systems on various protocols such as TCP, JavaRMI, and UDP. The work was publish as a demo paper at IEEE INFOCOMP Conference
As a part of Koc University Software engineering & design course, I worked with a team of five students to develop an invader game. The project was meant to practice various design patters such as Decorator, Controller, Observor, and Visitor.KuVid was awarded the best project of the year award
HuaweiQA is an android application that enables teachers to create quizzes for students and track their overall scores. The application was meant to showcase using HMS Toolkit, and I received first place in Huawei Android Development Competition.
About Me!
Hello there! I am Moayed Haji Ali,
A first-year PhD student at Rice University, working on multimodal generative modeling under the supervision of Prof. Vicente Roman. I recently graduated from Koc University, ranking third in the computer science department. I was also an undergraduate research assistant at KUIS AI lab. I am a proficient programmer with +6 years of programming experience and two year of industrial experience in machine learning. I strive to become a successful researcher and gain futher hands-on experience within the field of my interest in machine learning pretaining to computer vision, generative models and autonomous driving.
I thrive in diverse teams and competitive environments that polish my skill set and ensure constant growth. I am currently seeking research internship opportunities.
Ranked 3rd in the Computer Scinece department CGPA: 3.96 out of 4.0 (Full merit-based scholarship from Al-Ghurair Foundation for Education). Relevant Coursework: Autonomous Driving, Advances in Deep Learning, Deep Unsupervised Learning, Deep Learning and Computer Vision, Introduction to Machine Learning, Natural Language Processing, Algorithms and Complexity, Data Structure & Algorithms, and Software Engineering Analysis and Design. Teaching Experience: Deep unsupervised learning, Advanced rogramming in JAVA, and Introduction to Prorgamming.
Publications
M. Haji-Ali , G. Balakrishnan, V. Ordonez "Beyond Square: Extending Pre-trained Diffusion Models for Arbitrary Aspect Ratios" Under Submission 2023
M. Haji-Ali , A. Bond, T. Birdal, D. Ceylan, L. Karacan, E. Erdem, and A. Erdem "VidStyleODE: Disentangled Video Editing via StyleGAN and NeuralODEs" Published at ICCV 2023
Y. H. Nazarabadi, M. Haji-Ali , and N. Nayal "Opera: Scalable Simulator for Distributed Systems" [Demo Paper] Published at IEEE INFOCOMP 2021
Achievements
Huawei Koc university mobile development competition (ranked 1st)
2021
Project of the year award in software engineering analysis and design course
2020
Top 100 worldwide in IEEEXtreme Programming Contest (ranked 2nd in Turkey).
2019
Damascus ACM College Programing Contest (ranked 1st for two years).
2015 & 2016
Honorable mention in Asia-Pacific Olympiad of Informatics (APIO)
2016
Professional experience
KUIS AI LAB Computer Vision Research Intern
Istanbul, Turkey
“Text-guided image manipulation”. (Supervised by Prof. Aykut Erdem)
July 2021 - July 2023
Leading an ongoing project on unsupervised disentangled image representation using multimodal neural networks.
Lead a research project on video modeling and manipulation using StyleGAN2 and Latent ODEs.
Implemented using Pytorch a baseline for text-guided image manipulation based on CLIP and TokenGAN
Remote Part-Time Machine Learning Engineer at CareX
California, USA
“Body vitals estimation from fignertip and face videos”
Feb 2021 - Jan 2022, April 2023 - Aug 2023
Implemented a Pytorch and Tensorflow framework interface for integrating, training, and visualizing deep learning models
Developed a reliable AI-based method for blood pressure estimation from fingertip videos
Contributed to HRV estimation from fingertip videos technology, which was deployed in CareX’s SDK
IUI LAB Resarch Intern
Istanbul, Turkey
“Automated feedback for school assignements” (Supervised by Prof. Metin Sezgin)
May 2020 - June 2021
Designed and implemented an unsupervised in Tensorflow a feedback system for sketch-based questions based on a novel transformer-based architecture
INDEPENDENT PROJECTS & SKILLS
Coolingo:
An Android bot for teaching children English interactively
Developed a video chat-bot that combines several SOTA models effectively: BlenderBot 2.0, Google T5, MakeItTalk, FastSpeech2.0, and GPT3
Deployed AI models on AWS lambda, EC2 instance, and a local server
Video Continuous Implicit Representation
As a part of “advances in deep learning” course project, I implemented a new architecture using Pytorch for representing videos as an implicit function parameterized by a transformer architecture.
Co-Founder of Yders:
A platform that empowers university students and links them with school students to give them private tutoring at reasonable prices.
1st place in Startup Roadshow Istanbul 2019.
Technical Skills & Tools:
Programming Languages And Tools: Python, C/C++, Java, Julia, MYSQL.
Advanced algorithms, Data Structure, Computer Networking, Machine learning, Deep learning, and the basics of Web development, and Android development.
Unix, LaTex, Web scraping, Git, LEGO Robotics, and Arduino.
Contact
Please reach out if you have reasrch internship opportunities, or just want a simple chat.
Feel free to contact me via the form below or at mh155@rice.edu