At ThreatMark we build trust and safety in the digital world through state-of-the-art behavioral profiling solutions supported by AI-based security platforms. Our products deliver payment protection, ensure user identity, and detects cyber threats for businesses around the world. We’re detecting when a customer’s behavior seems out of character, blocking fraud in real-time – helping organizations outsmart criminals.

We’ve had consecutive years of outstanding growth and commercial success, and our team is enthusiastic, dynamic, and ambitious. Our mission is to make ThreatMark’s technology the No.1 risk management system globally.

We are currently building a dedicated Machine Learning Team that will work on hard problems in all phases from data collection, research to prototyping to enable product teams to construct and operate their production services as effectively as possible.

We are looking for skilled developers with the right attitude to push the boundaries of the status quo and create new technology.

Some example problems:

• Verify the identity of the person using data from mobile sensors/keyboard and mouse

• Detect the presence of financial malware in the browser memory

• Spot session takeover by evaluating navigation patterns

• Track users without using third party cookies

• Combine a lot of imprecise signals in the single probability distribution

• Do you feel excited thinking about how to deal with such problems?

• Did you write malware or defaced poorly secured websites as a kid? You are our guy!

• Do you go to any depth to discover the issue, even if it means debugging syscalls or low-level instructions? You are our guy!

• Are you rejecting any superficial explanations and go deeper to understand which exact numbers undergo some operation in crypto/ml and why? You are our guy!

• Do you laugh at code and tell yourself – I would use bloom filter here… This will lead to catastrophic backtracking in this regex. I would use Kalman filter to combine those readings, not average… This is a solved problem – Linus used Merkle trees for that in git – you are our guy!

If you prefer to wear a suit and not a dirty t-shirt, there is a more structured description for you.

What you’ll do:

As a Senior Engineer in the Machine Learning Team, you will build models, libraries and tools to be used in our products to further improve efficacy

• Make fast prototypes
• Build machine learning models
• Build python libraries which will be further used by the product team
• Design/Implement scalable, low-latency, high-throughput, fault-tolerant, extensible, and easily maintainable data processing pipelines for real-time systems.

What skills you’ll need:
We expect you to have those or to rapidly acquire them

• Programming (anything Turing complete, but mostly Python and JavaScript)
• Computer science skills (algorithms and structures, complexity, information theory, etc) at university level (we care about the skills, not the degree)
• Machine learning skills (data preprocessing, visualization, cleanup, feature extraction, model evaluation, etc…)
• Deep learning skills (MLP, RNN, CNN…)
• Statistics (Bayesian statistics, probability programming)
• Strong debugging, testing, tuning, and problem-solving skills
• Knowledge of IT world – HTTP protocol, networking, etc.
• Cloud engineering (AWS)
• Work with large datasets – they might be in the relational database, dump, proto-buffers, or you might have to collect it yourself

What would make you a strong fit:

• 5+ years of professional software development experience
• Demonstrable track record of exceptional software engineering skills on past projects
• Experience building highly available low-latency systems using any programming language. Strong familiarity with Python.
• Experience working with large datasets and best in class data processing technologies for both stream and batch processing
• Strong communication & collaboration skills
• Self-starter, with a quick learning curve.
• Knowledge of cryptography, IT security principles, attacks, malware