Milad Soltany

About Milad

Data Scientist and Machine Learning Engineer

Experienced Data Scientist and Machine Learning researcher and educator specializing in Computer Vision, Autonomous Vehicles, and Data Science with a strong academic background and extensive teaching experience.

I love to teach!

Since 2018, I’ve had the privilege of teaching over 2,000 students how to code in Python, guiding them through the fundamentals and advanced concepts of programming. Along with coding, I’ve introduced them to the world of Data Science, Machine Learning, and Deep Learning, helping them build a strong foundation in these rapidly growing fields. My goal has always been to empower students to bring their ideas to life through practical, hands-on learning and to inspire them to explore the vast potential of these technologies.

Machine Learning Researcher

My work has garnered over 180 citations, reflecting my commitment to advancing knowledge in machine learning, data science, and deep learning through research and education. I also have a Master's degree in AI from Queen's University in Canada =)

Experiences

Explore my innovative projects in machine learning and data science.

MASc in AI - Queen's University

Sep 22 - Current | Kingston, ON

• Currently working on tackling the challenges in Supervised Federated Domain Generalization (FDG) including client heterogeneity as well as inherent model overconfidence.

• Introduced a model for Unsupervised Federated Domain Generalization and established a theoretical link between domain shift and gradient alignment. Demonstrated superior performance over baselines by aligning gradients at both client and server levels for improved model generalization.

• Developed TimeHUT, a novel approach for time-series representation learning that dynamically adjusts temperature scheduling to balance uniformity and clustering, resulting in improved performance in time-series classification and anomaly detection.

• Utilized Queen's HPC and Compute Canada HPC resources to train complex models, ensuring efficient computation and accelerated training times.

• Managed the GPU clusters in the AIIM lab, ensuring optimal performance and availability for research projects through hardware maintenance, software updates, and resource allocation. Developed scheduling policies to maximize utilization and provided technical support and training to lab members.

Instructor at Educate360

Feb 23, Aug 24 | Remote based in Boston, MA

• Worked as an independent contractor to deliver personalized courses to high-end organizations

• Taught multiple courses, including Introduction to Python, Python for Data Science, and Python for Machine Learning using Numpy, Pandas, Matplotlib, Seaborn and Scikit-learn.

• Educated over 150 students in Python and Data Science and Machine Learning from diverse professional backgrounds remotely through online classes based on ZOOM.

GRA at Queen's University AutoDrive

Jan 23, Jun 24 | Kingston, ON

• Supervised the perception team, providing guidance and conducting weekly meetings to ensure alignment and progress toward competition goals.

• Operated as the GRA for the group and developed the necessary modules for the perception challenge including object detection, sensor fusion and distance measurement systems using cameras and lidars

• Worked closely with the Queen's AutoDrive team in the 2nd and 3rd year of international AutoDrive challenge

Lead Computer Vision Engineer and Data Scientist at ISAA

[Also known as Zibanegar Hoosh-Afarid]

Nov 21, Sep 22 | Remote

• Designed a digital smile system using diffusion models as part of a collaboration with ISAA.

• Optimized and deployed generative models on Google Cloud Platform GCP to ensure scalability and high performance.

• Utilized SQL to query and manage large datasets, enabling efficient data analysis and model training.

• Conducted weekly meetings to monitor team progress and address any challenges.

• Collaborated closely with the clients to understand their needs and translate requirements into technical solutions.

• Contributed to a US patent publication in 2022 as a result of this work.

Co-Founder, Instructor and Mentor at AIR Center and Pandas group

• Co-founded Artificial Intelligence (AIR) and Robotics Center

• Taught Python, Machine Learning and Data Science, and Deep Learning to over 2000 online students

• Collaborated to design Python, Machine Learning and Data Science, and Deep Learning courses to be taught at AIR.

• Taught machine learning and datascience tools and libraries including Numpy, Pandas, Matplotlib, Scikit-Learn, PyTorch and TensorFlow

• Achieved 1st team in the National Rahneshan Competitions for Autonomous Vehicles (Jan 2021)

• Achieved 2nd best team in FIRA Competitions in Autonomous Cars League (Race Section)

• Achieved 3rd best team in FIRA Competitions in Autonomous Cars League (Urban Section)

• Ranked 2nd best team in the national Data Days competitions among more than 140 teams by designing an ad-recommender system

• Helped design a fully autonomous vehicle based on AVIS engine for the FIRA competitions

• Collaborated in designing a platform for automatic parallel parking, path planning, path tracking and control

July 20, Sep 22 | Tehran