Machine learning engineer

Veridium is the global leader in mobile biometrics authentication, protecting the identity and data of customers and enterprises all from the palm of their hands, defending the industry and consumers against providers of fraudulent biometric technology.

In an increasingly digital world, with more devices and platforms connected to the Internet of Things, the need for ironclad security is paramount. Using your biometrics – your unique traits or behavioural characteristics – to prove identity will safeguard your company’s most critical assets in a way that’s both convenient and secure.

Today more than ever, people access their health and financial records, corporate documents, even conduct high-dollar transactions from mobile devices, making them vulnerable to breaches and hacking.

Everyone acknowledges that passwords are a weak link in enterprise security. You can lose them, share them, and crack them. Biometric authentication can strengthen legacy systems by adding an additional layer of security. With our technology, a company can deploy biometrics to replace tokens or passwords altogether.

YOU become the password using biometrics — your face, fingerprints, voice — to keep data untouched, private and convenient to access, send, receive and store.

We developed the global standard for biometrics-based identity authentication, called BOPS, the Biometric Open Protocol Standard. We contributed the BOPS protocol to the IEEE and it is now the approved international standard for biometric-identity assertion. With a BOPS compliant framework, we offer the only end-to-end, enterprise-grade platform that validates those accessing data are who they say they are.

BS.MS in Computer Science or a related technical discipline, and Master in AI

Having as many of the following skills represents an advantage:
- Knowledge of different types of machine learning algorithms (SVM, Kernel Ridge Regression, Random Forest, PCA, k-means, etc.) and know how to use them in practice
- Good understanding of neural network theory and how to train, test and evaluate modern architectures, e.g. convolutional neural networks, recurrent neural networks, auto-encoders, etc.
- Ability to combining different types of models and architectures in order to improve the performance.
- Knowledge of tuning models' parameters, e.g. using grid search.
- Feature engineering. Being able to understand the data that you are working with and extract useful features.
- Statistics and probability: Know how to evaluate a model or a solution in terms accuracy, precision, recall or other performance metrics. Understand probabilistic models like Naive Bayes , Hidden Markov Models, ROC curves, etc
- Signal processing: Use different types of signal processing methods for extracting relevant features, e.g. Discrete Fourier Transform.
- Read and understand different scientific papers in order to find ideas that can be applied to our solution.
- Being able to implement the solution in a distributed environment, e.g. Spark and Kafka.
- Knowledge of python and working experience with libraries such as tensorflow, keras, scikit-learn, numpy, pandas.

Research and develop a solution for user behaviour analysis

About the company
Veridium is a Digital Infrastructure Security Company devoted to developing highly secured Identity Assertion Platforms.
Headquartered in New York City, research and development labs operate in Bucharest (research machine learning, backend, mobile, QA, devops), Oxford and Boston ( computer vision).