Data Scientist (Fraud)

About the company

Kasada builds technology to stop malicious bots and automated abuse, helping protect online users and the brands they interact with. The company focuses on identifying and blocking harmful automation from the first request, including threats that have never been encountered before, and operates globally while defending some of the best-known organisations.

The team values curiosity, bold thinking, and a hands-on approach to solving tough cybersecurity problems.

Role overview

Kasada is hiring a Data Scientist to join the Account Intelligence team and help defend customers against fraud and malicious automation. This is the first data science hire for the team, and the work will centre on predictive modelling, statistical analysis, and production-ready detection methods that can hold up against real attackers.

You will stay technical across the full model lifecycle: exploring data, building and training models, evaluating their performance, and monitoring how they behave as adversaries change their tactics. You will work closely with engineering, research, and security operations to convert data science experiments into live defences.

What you'll do

• Design and improve predictive defences that support risk mitigation while protecting the user experience.
• Build, train, and iterate on models that can perform effectively against real adversaries at scale.
• Test models thoroughly and assess trade-offs so decisions can withstand review from engineering, product, and security teams.
• Collaborate with engineers, researchers, and product managers to move models from experimentation into production and keep refining them as attackers and customer needs evolve.
• Translate model outputs into clear, understandable explanations for both technical and non-technical stakeholders.
• Analyse large datasets to spot anomalies, reveal adversarial behaviour, and identify new attack patterns.
• Stay informed about adversarial machine learning and cybersecurity trends, and use relevant methods to strengthen detection capabilities.

What we're looking for

• Strong curiosity about fraud, attacker behaviour, false positives, and the real-world impact of detection systems on legitimate users.
• Comfort working across engineering, research, and product functions, with the ability to explain model behaviour and limitations clearly.
• At least 2 years of professional experience in data science or applied machine learning.
• Good grounding in statistics, sampling, time-series data, and practical predictive modelling such as gradient-boosted trees, random forests, or deep learning.
• Strong working knowledge of Python, SQL, and common ML libraries such as scikit-learn, PyTorch, and TensorFlow.
• Experience evaluating models in production or production-like environments, including precision, recall, calibration, and behaviour under adaptation.
• Excellent analytical and problem-solving ability, with careful attention to detail and a habit of pressure-testing your own work.
• Experience working in cloud environments such as AWS.

Bonus experience

• Background in fraud, trust and safety, account takeover, or abuse detection.
• Experience building or supporting machine learning systems in adversarial settings where attackers actively try to evade detection.

Tools and environment

The role works with AWS, ClickHouse, Python, scikit-learn, PyTorch, and TensorFlow.

Benefits

• Equity or stock options as part of the company’s global success.
• Support for growing families, including generous parental leave and pre-, during-, and post-leave resources.
• Wellbeing support, including access to an employee assistance program with confidential counselling for employees and their loved ones.
• Birthday leave.
• Wellness leave.
• Annual company offsites for collaboration and celebration.
• A dog-friendly headquarters in Sydney.

Application process

Applicants can expect a confidential, exploratory conversation with the team as the next step. The hiring process is intended to be thoughtful, efficient, and transparent.

Candidates may share their pronouns and any interview adjustments they need so the team can support them effectively.

Equal opportunity note

Kasada encourages applicants from underrepresented backgrounds to apply even if they do not meet every requirement, and values diverse perspectives and different career paths.

Additional note

Kasada is an E-Verify employer for US-based applicants only.