Our Team and Experience

Decades ago a circle of inspired researchers invented computers — now we are giving them eyes.
The mission
Every once in a while, a revolutionary technology comes along that changes everything. Mainframes, PCs, Internet, Mobile — these tech paradigm shifts had transformed the world forever. We are on a verge of a next technical revolution.

Spatial computing, Machine Learning and Autonomous Robotics will change for the better a way people create, work and live, creating unique opportunities for the humanity.

It's a great moment to witness an emerge of these new technologies. We are lucky enough to have a privilege in defining these cutting edge markets.

Dioram aims on becoming the key player in global SLAM market, creating a significant value for hardware manufacturers and vendors.
Oleg and other young scientists started weekly workshops to discuss Artificial Intelligence and Computer Vision problems and opportunities. The technical core of the team was formed.
2015
The History of the Future
In 2014 Oleg Scherbakov participated in a Computer Vision project of autonomous home robot. Oleg realised that existed SLAM and positional tracking solutions significantly lacked either performance, speed or robustness. It inspired him to look for a new approach of solving SLAM problems. This eventually led to Dioram.
In 2016 Oleg met Alex and they started collaboration to create an inside-out tracking system for VR. First preliminary exploratory and market research was carried out.
2016
The first version of visual odometry Core-1 was developed and R&D with IMUs was started. The SLAM-solution was set for an optical flow reuse.
2017
Alpha version of visual inertial odometry released. Started full-scale SLAM One development leading to Core-2 finalization. DeepTrack branch launched, R&D with Machine Learning usage for SLAM.

2018
The Core-3 of SLAM One platform was developed. New AR/AR hardware prototypes released.

Going off the stealth mode! Starting full-scale business development.
2019

The History of the Future

In 2014 Oleg Scherbakov participated in a Computer Vision project of autonomous home robot. Oleg realised that existed SLAM and positional tracking solutions significantly lacked either performance, speed or robustness. It inspired him to look for a new approach of solving SLAM problems. This eventually led to Dioram.
The science team
Dioram is honored to employ some of the best scientists and engineers in the industry. Many of the R&D team are
ITMO University
alumnus — one of the top Russian IT Universities which is well known for being 7th time winner of ACM International Collegiate Programming Contest.
Alex
Mentor and co-founder
Alex graduated with distinction from Polytechnical University.

He's a serial IT entrepreneur who founded more than 14 companies. One of his most prominent is eLama which is a top Google EMEA partner with more than $300m annual revenue.

Alex is passionate about science and AR/VR.
In additional to Dioram he founded Varwin — an Enterprise focused AR/VR cloud platform
Innokentiy
Deep Track developer
Innokentiy is an early Dioram contributor and collaborated with Oleg for the domestic robots Computer Vision project. He's the head of Neural SLAM tech branch.

As a Master of Science in Computer Photonics he's deeply interested in SLAM, semantic segmentation, FEM, SfM, Image processing, Deep and invariance learning, hypernets, AGI. He's the active university teacher of AI.
Ivan
SLAM One developer
Ivan is specialized in deep convolutional neural networks, image classification, feature-based image processing, symmetry detection, motion estimation. He's a Master of Science and did Lucas-Kanade Optical flow-based visual SLAM as his graduating project.

He's other projects include facial regions detection segmentation via deep convolutional networks, SLAM map history protocol specification for Unity and Ivan implemented optical motion detection module. He enjoys game-dev in his spare time.
Jaafar
SLAM One developer
Jaafar is a mechatronics engineer who specialized in perception and vision algorithms for Mobile Robots. He has a Master's degree in Intellectual robotics in ITMO, and now he is a post-graduate student in Control Systems & Robotics faculty and a researcher in AI for Cyber-Physical Systems Laboratory in ITMO University.

Jaafar's scientific interests are algorithmic thinking, programming, and mathematics. He participated in regional competitive programming contests and got several achievements in this field.
Kate
SLAM One developer
Kate is committed to the AR future receiving her Bachelor of Science graduation for virtual objects rendering and occlusion enhancement methods research. She's proficient in deep convolutional neural networks, autoencoders, SLAM, and feature-based image processing.

Kate enjoys dealing with Functional Programming languages and GPU-oriented parallel computing algorithms. She also did autoencoder-based learning algorithm implementation for rotationally invariant feature construction.
Kirill
Deep Track developer
Kirill graduated and received his Bachelor of Science degree in Electronics, radio engineering and communication systems in University of Aerospace Instrumentation(SUAI). Apart from Computer Vision and Visual Inertial Odometry Kirill is into signal filtration, diffraction, electromagnetic compatibility and inertial measurement units(IMUs).

He's is one of the biggest contributors to Dioram IMU-related advances creating methods for tracking and filtration, calibration and research of multi-IMU systems.
Nick
CTO
Nick is the one who ensures day-to-day smooth R&D operations and production. He worked with CISCO, Singularity.NET as a research engineer and several industrial projects like CNC machines and mechatronics. Nick's primary interests are visual odometry, object detection, SLAM, deep learning, structural and syntactic pattern recognition, generative models, symbolic AI, mechatronics, metaprogramming and a Compiler Theory.

Nick is a contributor to RE-ID, Semantic Image Retrieval and Invariance learning projects. He's an aviation technology and avionics hobbyist.
Oleg
Chief Scientist and co-founder
Oleg Ph.D. is the founding member of Dioram and has got an astonishing academics experience. His professional interests are AGI, image analysis, game theory, Deep Learning, RL, Structure Learning, Manifold Learning, Deep learning based CV, SLAM, VIO.

Prior founding Dioram he collaborated with several enterprise projects including domestic robots and a blockchain technology. Oleg is an assistant professor and a lecturer at ITMO university. He's a author of numerous scientific papers and gives scientific conferences talks.
Paul
Elektronic developer
Paul recently graduated with Master of Science degree in Computer Science(ITMO University) with a work of autonomous vehicles trainable models. Paul's scientific interests are cognitive modeling, probabilistic programming, deep reinforcement learning, reservoir computing, temporal sequence and dynamic behavior learning, semantic segmentation, instance segmentation.

Currently he's doing autonomous driving agent for video game "Grand Theft Auto V" powered by deep learning.
Serge
Engineer
Serge's got more than 13 years of hardware engineering experience working with a large enterprise producing precise measurements devices. He designed, produced and tinkered numerous devices and devkits for work and leisure.

He has been a radioelectronics fan since childhood. Serge enjoys developing in C for embedding devices and microcontrollers.
Vas
CEO
Vas is a highly experienced executive with a more than 12 years in the industry.

His passion is emerging tech and he launched products in many markets including mobile, fintech, social networks, and more. Vas is a long-time AR/VR enthusiast, evangelist and visionary.

Dioram team scientific developments

The team members brought all their passion and experience to the project. As the science company Dioram relies heavily on past and future achievements of the team
Awards and Recognition
Among scientifics awards received by Dioram team are:
  • Kurzweil Best AGI Idea Prize 2018 (for the paper "Semantic Image Retrieval by Uniting Deep Neural Networks and Cognitive Architectures")
  • Kurzweil Best Paper Prize 2018 (for the paper "Vision System for AGI: Problems and Directions")
  • Best Bachelor of Science project implementation award, ITMO University, 2017
  • Best paper at "Modern problems of radio electronics" scientific and technical conference, 2017. (for the paper "Methods of increasing the capacity of networks of fixed radio access with code channel separation")
Professional recognition is a important part of self-fulfillment. That's why many team members have been giving speeches at professional conferences or scientific talks. They include:
  • International Conference on Artificial Neural Networks (Rodos, Greece), 2018
  • "Science, Training and Doctrine" conferences of ITMO University 2017-2019
  • 16th International Conference on Next Generation Wired/Wireless Advanced Networks and Systems(St.Petersburg, Russia), 2016
  • Communication and radio navigation systems (Krasnoyarsk, Russia), 2016
  • International Conference on Future Communication Technologies and Engineering (Shenzhen, China), 2014
  • 11th International Conference on Quality Control by Artificial Vision (Fukuoka, Japan), 2013
  • International Conference on Machine Vision (London, UK), 2013
There are several guest speeches at international science conferences coming from Dioram team in 2019 — stay tuned for updates!
Publications
There are more than 50 of scientific publications made by Dioram team members and contributors. We are happy to share some of our advances in Dioram blog. More publications from Dioram are to come in 2019! Dioram papers include:
  • Hypernets for ensemble learning, 2019
  • Semi-automatic road segmentation via fully coordinate convolutional deep neural networks, 2019
  • Symmetry line detection methods research for photopletizmography systems, 2019
  • Vision System for AGI: Problems and Directions, 2018
  • Semantic Image Retrieval by Uniting Deep Neural Networks and Cognitive Architectures, 2018
  • HyperNets and their application to learning spatial transformations, 2018
  • Research of criteria for the assessment of the intra-system electromagnetic compatibility of the wireless communication system, 2017
  • Method of improving the spectral efficiency of telecommunication systems based on the approximation of the bending signals with atomica functions, 2017
  • Investigation of Questions of Non-harmonic Signal Scattering on Impedance Structures, 2016
  • Evaluation of the In-System Electromagnetic Compatibility of a Radio Communication Network with Code Division Channeling, 2016
  • Analysis of the passage capacity of the wave-leading cjmmunication channel of EHF range, 2016
  • Solving of an erythrometry problem using a specialized Hough transform 2015
  • Modifying the Hough transform by using the periodicity of the regenerated structure of objects on an image, 2015
  • A convolutional autoencoder as a generative model of images for problems of distinguishing attributes and restoring images in missing regions, 2015
  • Fast reconstruction of Go board grids using the modified Hough transform, 2015
  • Image inpainting based on stacked autoencoders, 2014
  • Curvature histogram features for retrieval of images of smooth 3D objects, 2014
  • Practical algorithmic probability: an image inpainting example, 2013
  • Erythrometry method based on a modified Hough transform, 2013
  • Method of prediction based on algorithmic probability in the problem of image restoration in missing regions, 2013
  • Hough and Fourier Transforms in the Task of Text Lines Detection, 2013
  • Geometrical normalization of three-dimensional biomedical images for efficient presentation and compression by means of octrees, 2012
Academics
Most Dioram of developers are from academics. Helping others to learn, carrying of experience and scientific methods from one generation to another are important parts of Dioram values. This is why the team members are found themself as active lecturers and teachers. These are ITMO and SUAI universities courses conducted by Dioram team:
  • Artificial Intelligence
  • Computer Vision
  • Machine Learning
  • Practical training on modeling in Matlab/Octave
  • Enterprise architecture
  • Game Theory
  • Applied Mathematics
  • Сommunication and telecommunication systems
  • Electrodynamics and propagation of radio waves
  • PEMS systems
Supervision: Master's double degree program supervision in collaboration with Lappeenranta-Lahti University of Technology, Finland. Prepared 5 B.S. and 4 M.S. students with specialization in Computer Science Conference Reviewing: Artificial General Intelligence (AGI-18), 2018, Artificial General Intelligence (AGI-17), 2017, Machine Vision Applications (MVA2017), 2017.

SLAM solution request

Complete the form and give us information about the SLAM tracking you require.
Ask us anything about Dioram SLAM One!