PaaS for research, deep-learning model training and ultra-convenient coding education environment.
Open-source edition for deploying / developing your own Backend.AI Server Farm.
Documents, forum, showcases of Backend.AI platform.
Proposes new paradigms for computing-based research via cloud computing and artificial intelligence (AI) technologies.
Lablup offers Backend.AI, a distributed execution management framework specialized for deep-learning model training and computational researches. We distribute Backend.AI both open-source (https://www.backend.ai) and cloud backend hosting service (https://cloud.backend.ai).
On top of Backend.AI, we build CodeOnWeb (https://www.codeonweb.com), computing-based research and education platform and use.backend.ai, machine-learning model streaming interface service.
Lablup is a three-year old startup in South Korea, membered seven Ph.D cofounders and industry/acdemia-leading associates.
Our key motivation is to fill the gap between cutting-edge innovations and their impacts to the society and people. As the complexity of research and engineering increases at an unprecedented pace, the humanity needs a revolutionary way of sharing research methods on standardized, scalable computing platforms. Therefore, we set our goal as to make computing ubiquitously available to researchers and engineers, regardless of which fields and divisions they work on. To achieve it, we decided to develop software products that accelerate the lifecycle of everybody's idea: research, development, sharing, collaboration, deployment, and public services.
Lablup began its voyage and created an early prototype of (formally known as) Sorna (now Backend.AI) (the distributed code execution engine) at Google Campus Seoul in May 2015. During early stages, we had created a variety of prototype products ranging from data-driven API-enabled IoT hardwares to machine-learning based chat-bots. Collecting feedbacks from our investors and mentors, Lablup has moved its focus to the code execution platform and a coding education service as its application. At the end of 2015, Lablup launched CodeOnWeb as the first product where people can learn and teach machine-learning and programming with zero-config coding sandboxes instantly available on web browsers.
Since 2016, Lablup has continued to enhance Sorna for app streaming. We have open-sourced Backend.AI (project Sorna) in November 2016 and launched the public Backend.AI API service on the clouds in January 2017. Backend.AI aims virtual machine learning / HPC cluster orchestration / management framework with various computing resources. Lablup officially launch Backend.AI 1.0 in Oct. 2017. Backend.AI also provides SDK, API streaming modules for 4 programming languages and plugins for various applications such as Visual Studio Code, Atom editor and IntelliJ (with PyCharm).
Now Lablup is upgrading Backend.AI to support various and heterogeneous resources with unified interface and scheduler. Backend.AI starts supporting TPU on Google Cloud Computer, enhancing super-computer scale orchestration with Nvidia DGX series, handling and load balancing various public cloud resources with Amazon cloud, Microsoft Azure and Google Cloud Platform as the one Backend.AI cloud.
Lablup is also testing Use.backend.AI service, human-friendly machine-learning service engine as a centralized gateway to streamed cloud machine learning applications.
Known as the lead developer of Textcube Project. As one of the Korea’s foremost experts on open-source projects, he has participated in many successful ventures for the past 10 years. Jeongkyu received his Ph.D in Physics, POSTECH with complex system and neuroscience studies.
He has been delved himself into the study of DNA-protein interactions using single-molecule optical microscopy techniques. Sharing the goal of improving research methodology, he changed his course of life. Jonghyun received his Ph.D. in Physics at POSTECH.
With full-stack engineering knowledge and skills, he has participated in both the back-end framework and front-end refinements of Textcube Project. Joongi developed a GPU-accelerated packet processing framework offering world-first 80 Gbps performance as his Ph.D research in Computer Science at KAIST.
She is a computational and systems neuroscientist who wants to read mind through oscillations in the brain. She is also interested in the mechanism(s) how juveniles learn from experience. She believes that the neuroscience and the artificial neural network can provide reciprocal insights into making better (understanding of the) world.
He has been a researcher in the field of chemical engineering for 10 yrs. Now, like a loading buoy helping tankers to load and offload oil, he wants to interconnect ML with other research areas.
Performing the role of Senior research Artificial Intelligence. Strong expertise in the High Performance Computing Area. Advanced cloud computing competitive intelligence skills and interests in machine learning, neural network, deep learning. Experienced enterprise service delivery provider.
Jaepil is an entrepreneur and a full-stack software engineer. He worked for 5+ startups for 10+ years - as a software engineer, product manager, tech lead, and CEO with a social network service, an automatic content recognition system, a e-commerce aggregation system and O2O/IoT platform.
Jungseung is backend service engineer who has strong experience of the orchestration for virtualization, network and server.
He has experiences in many exciting domains of computer programming, especially network applications with security and performance, and has participated in many companies developing sms-gateway, network security, social networking platforms, mobile game servers and embeded devices.
We are using Backend.AI as the core of Backend.AI Cloud service and APIs for interested groups of individuals and companies.
Cloud-based code execution service for research, deep-learning model training and ultra-convenient coding education environment. Provides instant access to various runtimes including popular research programming languages such as Python, Julia, R as well as machin-learning toolkits like Tensorflow and Caffe.
Open-source Source code for deploying / developing your own Backend.AI Cloud Server Farm. You can use full potential of Backend.AI with your own on-premise servers, including zero-configuration deploy, autoscaling, dynamic agent registration and IDE integrations. We are also working on combination feature with heterogeneous cloud services.
Web-based platform for learning & teaching programming, with user-created lecture courses and instant-access coding sandboxes.
Join the open beta at codeonweb.com now.
An education platform aimed for computational thinking along with related STEM (science, technology, engineering, and mathematics) knowledge sets with ultra-personalized coursework system.
Discovers the potential of computational thinking from kids by teaching Python and related elementary school maths/sciences.
Lablup Inc. : Make AI Accessible
Proposes new paradigms for
computing-based research via
cloud computing and
artificial intelligence (AI) technologies.