Contact person: AI-Based Cybersecurity Adversarial Attack Detection -- 2

Phone:Show

Email:Show

Location: Sydney, United Arab Emirates

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"Key Responsibilities:

- Conduct in-depth literature reviews on state-of-the-art AI-based adversarial attack detection algorithms in cybersecurity, with attention to current trends, limitations, and future directions.
-Formulate AI algorithms, then build and train models targeting a specific class of cyberattacks using cutting-edge AI and deep learning techniques.
-Search, identify, and preprocess suitable adversarial attack detection datasets, applying best practices in data mining, feature engineering, and effectively managing imbalanced/noisy data.
-Run experiments on a GPU-powered AI server, ensuring efficient model training through optimal configuration, hyperparameter tuning, and effceint evaluation metrics.
-Contribute to academic publication writing and (optionally assist in patent documentation and filing, depending on the novelty and applicability of the developed solution).
-Support technical documentation and presentation preparation to communicate progress and results to stakeholders throughout the project.
-Design and implement a functional application (web or cross-platform; to be developed in a later phase under advisory guidance) that showcases real-world deployment of the developed AI models.

Preferred Qualifications:

-Bachelor’s (final year), Master’s, or PhD-level background in Computer Science, Cybersecurity, Artificial Intelligence, Machine Learning, or a closely related field.
-Strong experience in Python programming and Cyber and AI libraries such as PyTorch, TensorFlow etc.
-Solid understanding of cybersecurity and AI theories.
-Hands-on experience with AI model development, data preprocessing, and GPU-based training and testing environments.
-Experience working with real-world cybersecurity datasets.
-Familiarity with tools for building web or desktop MVPs is a plus.
-Strong academic writing skills and able to co-author high research outcomes to produce quality research papers and technical reports.
-Plus: Prior exposure to research publication, conference submissions, or patent drafting.

Deliverables:

-Documented literature review and comparative study of adversarial attack detection methods.
-Custom theoretically developed and built AI algorithms trained on curated datasets, including source code and performance reports.
-Cleaned and preprocessed datasets with applied feature engineering techniques.
-Trained models with reproducible experimental setup and tuning logs, deployed via GPU infrastructure.
-A functional application (web/desktop or on Edge computing) demonstrating use of the AI models in a real-world scenario (developed in the later project phase).
-Contributions to co-authored academic publication(s) and optionally patent draft documentation.
-Complete technical documentation and support material for progress presentations and future development phases.

Project duration: up to 6 months" (client-provided description)