About the job
Some of the challenges are to detect physical and digital forgeries, extract textual and visual data from identity documents, evaluate photo content and quality, detect user impersonation attempts, verify legitimate users through facial recognition, and make complex verification data easy for our customers to use and understand.
You will take ownership of document fraud R&D and be responsible for developing document anti-spoofing methods through all process stages until a new feature is deployed into the product.
This is a key role helping our product team as an applied scientist with a strategic mind and high research orientation, and first line of developing document anti-spoofing features.
This role would require you to do the following on a daily basis:
Formulate innovative approaches to combat document fraud and reduce risk with the powers of ML.
Create new features, train new models, deploy them into production environment.
Contribute by extending and improving our ML frameworks and platform, creating next-generation capabilities.
Build and deploy solutions to interesting computer vision or machine learning problems including document data extraction, fraud detection or biometric verification challenges.
Support and guide other engineers in learning about, applying and delivering product features driven by machine learning techniques.
Work alongside other machine learning and computer vision specialists in order to deliver on both short term objectives and long term goals.
Help develop robust model training and data infrastructure to support continual optimisation of ML-driven approaches.
Assist in steering the ML-led development across the tech team.
Gathering of information about new tech trends from academic articles, journals, code repositories…
About you
The ideal person for this role:
PhD degree in Computer Science (or related quantitative field) or MS degree in Computer Science with related experience.
5+ years of experience building machine learning systems in production, and with real-time technology problems.
Strong academic and publication record.
Experience with cloud-based training and deployment pipelines.
Experience training Neural Net architectures for classification, object detection, and segmentation.
Excellent coding skills (Python is essential, C++ is considered a plus).
Proficiency with some of these machine vision and machine learning frameworks: OpenCV, TensorFlow + Keras, TensorRT, PyTorch, Pytorch + FastAI. With a strong portfolio of development examples;
Transformers management framework, such as BERT, is a plus.
Good working knowledge of the tools in our dev stack, including Git, Google AI Cloud, Docker, and Kubernetes. A plus Linux, Redis, and ELK stack.
Solid understanding of statistics, probability, linear algebra & calculus.
Hands on experience working on computer vision and machine learning projects e.g. face verification, object detection and/or classification.
Comfortable reading, discussing, and applying research from published papers.
Communication is important so we expect you to be able to translate complex ideas into understandable content.
A pro-active, self-managing attitude.
What we offer in return
Learning days. You can learn during working hours.
We encourage the dissemination of knowledge both through internal meetings and by sharing our experiences with the community. Feel free to propose talks, open spaces, workshops, …
Training budget for personal and team formation.
Free day your birthday.
A flexible working environment. Currently due to COVID, it is fully remote but for the right person, it can be fully remote regardless. However, we do require you to be in our office from time to time if needed. Note that this position cannot be fully remote for every candidate and the decision will be made based on the candidate.
Competitive base salary. Additional year end bonus can be offered based on individual performance and company performance.
About ALiCE
ALiCE is a biometric identity verification solution that allows the online onboarding of new clients, reducing identity fraud and maximizing conversion rate. ALiCE offers a frictionless user’s identity verification in a two-step process: user takes a selfie and captures his ID card, ALiCE does the rest.
ALiCE Biometrics, as a spin-off from the R&D Technology Center Gradiant, was born with the mission of developing the best-in-class onboarding identity verification solution that uses Deep-Learning based Face Recognition and Passive Liveness Detection technology.
We use a lot of exciting technology. This is our technology stack:
Python for our service back-end code.
React, Swift, Kotlin and Javascript (based on Vue.js framework) for front-end work.
RabbitMQ and ELK stack for events queue management, observability and visual representation.
Domain Driven Design as main principle to domain modeling and keep focus on the product.
Test Driven Development to encourage the Outside-In design and improve the quality of our code.
Github for repositories management.
Github Actions for Continuous Integration and Continuous Deployment.
Notion for project management and documentation.
Kubernetes, Docker and Helm to orchestrate our services.
Google AI Cloud and Kong for underlying infrastructure.
Know more about our culture and challenge, please visit our web: https://alicebiometrics.com/jobs