Why Invest in Machine Learning (ML) Hub?
Easy-to-use, exclusive 🤗HuggingFace as a Marketplace.
Home of exclusive AI
The short video pitching our idea.
You can visit the ML-Hub.
What is the problem
ML-Hub allows us not only to monetize the idle pre-trained models but also to host only state of the art models by providing a competitive environment so it saves you from getting lost among models.
We bring the world’s best (accuracy + speed + security + users’ feedback) AI models together and make them available to developers/enterprises in one place. We will make it happen with the exclusive models trained by not only the ML-hub team but also our community.
Our marketplace allows developers to discover, test, and integrate various AI related APIs. Developers can see the demo of the API, play with them right from their browser, subscribe to the API, and then integrate one by copying a simple code snippet.
|Business Values||HUGGING FACE||ML-HUB|
|Do model uploaders earn money from their exclusive SOTA models?|
|Do users seek out the best models fast and easily?|
Hard to choose right one among overcrowded models
Only proven models accepted
|Are there sufficient business-tailored models to use?|
Lack of domain-specific models such as Healthcare, Legal and Finance.
Primary focus area : Domain-specific SOTA models
Business-tailored exclusive models
- Quality of our AI models comes from both the wisely-use of transfer learning and also data labeling capability.
For example :
De-identification of Personal Health Information (PHI) from medical records is one of the highly-challenging domain-specific NER (Named Entity Recognition) tasks. Our model outperforms current SOTA by %10 on the i2b2 official de-identification test data.
Unique infrastructure for model training and inference.
- Our infra is established on exclusive ML-Ops technologies such as Kubernetes, Kubeflow, Kserve.
- AI models will be able to run not only on the desired cloud platform but also on customers’ private servers.
- We offer minimum latency and scalability of model inference.
- Thanks to our unique infra, low prices will become inevitable.
We will save you from getting lost among thousands of AI models on the market.
- Based on the leaderboard approach we will develop, only a certain number of SOTA AI models will be able to serve. In this way, we will both keep the quality at the highest level by encouraging competition and also the idle AI models in the repos will make money.
To become the market leader with a 30% market share in the domain-specific AI Marketplace in two years.
Who We are?
J.Murat G.- MSc
Russell C.- MSc
ML Engineer Lead
Joe B. - PhD (Industrial Engineering)
Armando R. - MSc
DevOps/MLOps Team Lead
NLP Engineer Lead
Bunyamin K.- PhD (Healthcare Management)
Annotation Team Lead
Zahire S.- PhD (Mathematics)
CV Engineer Lead
Data Engineer Lead
Software Engineer Lead
Serdar Gümüs - PhD (Mathematics)
Dev. Sec. Ops. at Ebay
Data Analyst at IBM
Dev.Ops.Software Engineer at Bank of America
Use of Funds
Our business model is based on the fees we receive from our customers based on both B2B and B2C