3D Face Reconstruction Computer Vision Algorithms Engineer

Salary
£30,000 - £50,000 per annum depending on experience
Hours
7 hours a day
Position Type
Full Time
Location
Woking, Surrey
Closing Date
01/11/2019
Advert expires on
20/09/2019

Atom Universe

Main Duties

We are building a small but experienced and highly effective team of data scientists and software engineers to work on an R&D project of reconstructing people faces as 3D models from a raw point cloud captured by a mobile phone camera. The primary focus of the project is to build a framework to jointly learn a nonlinear face model from a diverse set of raw 3D scan databases and establish dense point-to-point correspondence among raw scans.



The Computer Vision Algorithms Engineer role is to design and build a framework for 3D face reconstruction based on CNN approach. The framework will treat input raw scans as unorganized point clouds to convert them to identity and expression feature representations, from which the decoder networks recover their 3D face shapes.


In this role you will have to deal with large-scale data set with intensive hands-on code development. In addition, this role will be responsible for the following: collect, process and cleanse raw data from a wide variety of sources. Transform and convert unstructured data set into structured data products. Identify, generate, and select modeling features from various data sets. Train and build machine learning models to meet quality and business goals. Innovate new machine learning techniques to address quality and accuracy needs. Analyze and evaluate performance results from model execution.


Are you deeply accountable for your work? Your passion for product ownership and track record of product development will prove critical to your success on our team. We are in need of a creative thinker with deep expertise in machine learning, algorithms, and optimization. You will work with amazing colleagues, brainstorm new ideas, and develop models and algorithms to solve challenging problems that have a substantial impact.


We are looking for a talented Computer Vision Engineer. The selected candidate will be doing R&D in image processing, computer vision, and machine learning tools for point cloud and image analysis, clustering, classification, image segmentation and anomaly detection. You will join us on projects that impact hundreds of millions of users.


Responsibilities


Contribute research that can be applied to AURA platform development


Collaborate with team members from prototyping through production


Analyze and improve efficiency, scalability and stability of various deployed systems


Conceive proof-of-concept prototypes that establish overall system performance


Apply machine learning to computer vision problems


Research, develop and prototype advanced software technologies related to tracking, 3D reconstruction, object detection, landmarks detection and appearance modelling


Develop novel, accurate computer vision algorithms and advanced systems with a focus on real-time face tracking, SLAM, and 3D face reconstruction from a raw and dense point cloud.

Skills and Experience

Key Qualifications


Passion for applying advanced methods, and innovating approaches at the intersection of machine learning, optimization, and computer science.


Self-learner and has a thirst for continuing learning with a passion for work, attention to detail, and a can-do attitude.


Prototyping skills


Understanding of applied mathematics and numerical optimization


Middleweight experience developing real-time computer vision software in C++, including algorithm design and systems software development


Currently has or is in the process of obtaining a PhD in Computer Science, Computer Vision, Machine Learning, Robotics or related technical field


Preferred Qualifications


Familiarity or experience with Unreal Engine is a plus


Writing high performance, memory efficient, and multi-threaded code


Highly skilled in C++


RGB-D 3D reconstruction algorithms


3D mesh algorithms, operations, texturing, and parameterizations


Hands-on experience with big data systems (e.g., MapReduce, Spark) with TB to PB scale datasets.


Coding skills and experience in Python based on state-of-the-art machine learning and neural network methodologies (e.g., TensorFlow, PyTorch) for training and serving.


Experience training deep convolutional and/or recurrent neural networks


Middleweight experience with proven track record with real-time object tracking, SLAM (Simultaneous Localization and Mapping), real-time image processing, etc.