Data Scientist

About Me

🚀 AI-ML Specialist (3+ years exp.) with a Master’s in Computer Vision. Experienced in developing AI solutions to real-world problems using Machine Learning, and Python, PyTorch, Tensorflow. Former Research Assistant at top tech institutions. Seeking roles in AI-ML, Data, and Applications.

I studied Computer Science in my bachelor degree at Nazarbayev University in Kazakhstan 🇰🇿. I did research in robotics 🤖 during my junior year and had my first paper published. I had summer internship at KAUST, Saudi Arabia in molecular visualization 🧬 and computer graphics. I participated in hackathons, did side projects and launched a React Native mobile app. I completed MSc in Computer Vision at MBZUAI in Abu Dhabi and successfully defended my thesis 👩🏻‍🎓. My thesis result was accepted as a conference paper to VISAPP 2023 (part of VISSIGRAPP).

Social Links

LinkedIn

Google Scholar

Technical Skills

Programming Languages: Python, C/C++, Java, JavaScript

Libraries: PyTorch, TensorFlow, pandas, NumPy, Matplotlib, Scikit-learn, Seaborn, OpenCV, NLTK, spaCy, Streamlit

Other: MLOPs, ETL, Git, GitHub, Linux, REST APIs, SQL, Microsoft Azure, HuggingFace, HTML5/CSS3, OOP, Data Structures & Algorithms, Excel

Key Words: Machine Learning, Deep Learning, Neural Networks, Natural Language Processing, Transformers, Generative AI, Data Visualization, Data Analysis

Computer Vision Tasks: Object detection, object recognition, classification, counting, segmentation

Machine Learning Tasks: SHAPley values, X-AI, time-series anomaly detection, classification, regression, clustering

Education

MSc in Computer Vision @ Mohamed bin Zayed University of Artificial Intelligence (Jan. 2021 – Jan. 2023)

• Fully-funded Scholarship, CGPA: 3.67/4.0
• Coursework: Mathematical Foundations of AI, Human and Computer Vision, Visual Object Recognition and Detection, Geometry for Computer Vision, Deep learning

BSc in Computer Science @ Nazarbayev University (Jul. 2016 – Jun. 2020)

• Fully-funded Scholarship, CGPA: 3.35/4.0
• Coursework: Software Engineering, Operating Systems, Algorithms, Artificial Intelligence, Database Systems, Machine Learning, Performance and Data Structures, System Design and Analysis.

Work Experience

Research Associate @ Mohamed bin Zayed University of Artificial Intelligence (September 2024 – Present)

• Support research of humanoid robotics on the human motion capturing, biomechanical computation, as well as the data science of human movement (2D/3D pose estimation, tracking)

Machine Learning Engineer Intern @ TachyHealth (_June 2024 – August 2024)

• Worked on Sustainability and Explainability for Fraud Detection in Healthcare Systems

Research Assistant @ Zayed University (June 2023 - September 2023)

• Composed a survey paper on privacy and trust issues in blockchain and edge computing resource allocation
• Collaborated with a team to analyze and present data from 50+ surveys, leading to new research insights

Machine Learning Engineer Intern (R&D) @ FortyGuard (June 2023 – July 2023)

• Assessed thermal comfort index datasets globally for indoor/outdoor environments
• Created user stories based on consumer needs and company product features

Research Assistant @ Mohamed bin Zayed University of Artificial Intelligence (February 2023 – July 2023)

• Built 3D detection Transformer model and trained on point cloud ScanNetv2 dataset
• Achieved significant improvement in small-size category detection: 3.51% (picture) and 1.4% (garbage bin)

Computer Vision Engineer Intern @ AD Ports Group, Maqta Gateway (October 2022 – December 2022)

• Worked on Automated Cargo Recognition and Counting System to improve the operational efficiency of the port
• Trained YOLOv5l model on custom dataset increasing accuracy from 95.9% to 98.8%

Machine Learning Engineer @ Nazarbayev University (January 2020 – January 2021)

• Created an end-to-end ensemble of Random Forest Classifiers for time-series anomaly detection in health data
• Obtained at least 79% accuracy and 20% false positive rate. Published the article in Internet of Things journal

Visiting Student Intern @ King Abdullah University of Science and Technology, NANOvis Group (May 2019 – November 2019)

• Worked on cell visualization and molecular graphics project
• Developed an automated traversal of a 3D scene of HIV virus structures (RNA, DNA, protein cells) using C++ and Qt framework

Front-End Software Engineer Intern @ KazDream Technologies LLC (February 2019 – March 2019)

• Implemented a login page, input validations, and a date-picker of the website using JavaScript, HTML5/CSS3, Vue.js. Logged outputs in GitLab, reviewed and fixed codes. Captured data from 3rd party APIs

Undergraduate Researcher @ Nazarbayev University, Robotics Lab (September 2018 – December 2018)

• Developed an Android tablet app in Java for medical staff to interact with a humanoid robot NAO to assist children with autism spectrum disorder. Established TCP connection between a tablet and Choregraphe software

Mobile Developer @ nFactorial School (June 2018 – August 2018)

• Created a React Native app SpeechArt(1000+ downloads on Play Store) for correcting speech defects with tongue-twisters and became one of 6 finalists out of 100

Projects

PDF Sentiment Analyzer | Python, Streamlit, tabula, pandas, NLTK, spaCy, HuggingFace, Transformers

Sentiment analysis web interface using VaderSentiment with text parsed from pdf file uploaded to local host

Fine-grained Classification | Python, PyTorch

Worked with three different datasets: Caltech’s CUB-200-2011, Stanford Dogs Dataset, and FoodX-251. Worked with models: ResNet34, ResNet50, ResNet101, Deit (Data-efficient transformer), ViT small, SEResNet50, NasNet. Proposed a novel model based on ResNetv2. Report link

Large-Scale Aerial Image Recognition | Python, PyTorch, Detectron2, Git

Fine-tuned Faster R-CNN using Detectron2 framework and increased mAP of small objects by 1.9% on a large-scale aerial iSAID dataset. Included skip-connections with upsampling for global feature extraction in FPN (feature pyramid network) and used shadowing augmentation to improve dataset quality. Report link

Blood Cells Classification and Localization | Python, PyTorch, Git, Yolov5

Adapted \emph{YOLOv5l} towards BCCD (Blood Cell Count and Detection) dataset and obtained 4.3% mAP on top of the baseline reducing inference speed by 1.7 ms. Solved the overlapping issue of red blood cells. Used copy-paste augmentation technique for fixing overfitting, focal loss for class imbalance reduction, and pre-trained Yolov5 model for better weight initialization. Report link

Efficient Human Pose Estimation | Python, PyTorch

Changed original Stacked Hourglass architecture by adding dilated and depthwise convolution filters in the ResNet Bottleneck. Improved the training speed by 25% on MPII Human Pose dataset

Retinal Eye Disease Detection with Convolutional Neural Networks

Experimented with various CNN networks such as LeNet, ResNet variants (ResNet32, ResNet50, ResNet101), VGG in PyTorch

Train Ticket Reservation | Java, SQL, MySQL, Git, GUI

Database Systems course project: desktop app for reserving tickets for a train seat using Java and SQL

Awards & Scholarships

• 1st place in Pioneers Industry 4.0 Data Science Hackathon, EDGE and MoIAT of UAE, November 2022
• 3rd place in ‘AI for Good’ Hackathon, Microsoft & NU, March 2019
• Big Data workshop by CS Prof. at Stanford, Jennifer Widom, NU, September 2019
• Finalist of Mobile App Dev. Bootcamp @ nFactorial School, June - August 2018
• Bertelsmann Data Science Scholarship program, Udacity, May 2018

Publications

• Nurakhmetova, A.; Lahoud, J. and Cholakkal, H. (2023). Data-Efficient Transformer-based 3D Object Detection. In Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP; SciTePress, pages 615-623. Link
• Yazici, A., Zhumabekova, D., Nurakhmetova, A., Yergaliyev, Zh., Yatbaz, H. Y., Makisheva, Z., Lewis, M. & Ever, E. (2023). A Smart e-Health Framework for Monitoring the Health of the Elderly and Disabled. Internet of Things, Volume 24. Link
• Sandygulova, A., Zhexenova, Z., Tleubayev, B., Nurakhmetova, A., Zhumabekova, D., Assylgali, I., Rzagaliyev, Y. \& Zhakenova, A. (2019). Interaction design and methodology of robot-assisted therapy for children with severe ASD and ADHD. Paladyn, Journal of Behavioral Robotics, 10(1), 330-345. Link

Certifications

• Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization (Coursera)
• Neural Networks and Deep Learning (Coursera)
• Bertelsmann Data Science Scholarship program (Udacity)
• Web Developer BootCamp (Udemy)
• Python Object Basics: Functions, Recursion, and Objects (Coursera)
• Python Data Structures (Coursera)