Frequently Asked Questions

 

  1. What is iVentura?

iVentura is a self-service platform for Analytics and Machine Learning that provides a collection of ML components and services natively running on Kubernetes and it is fully configured development environment most flexible way to do Machine learning, Deep learning on the cloud and on-premises.

2. What types of notebooks are supported?

Presently, Jupyter notebooks such as python and R are supported.

3. How do Notebooks work?

With Notebooks, you can register the account and start working with notebooks within a few seconds. Notebooks give you access to all features, such as distributed training, hosting, and experiment management.

 4. How to import the libraries? / How to install additional dependencies?

Choose the terminal catalog in iVentura workspace and install the additional libraries and dependencies for your model.

For instance, # pip install NumPy, # conda install fbprophet.

5. What business case a potential customer will solve by subscription of the platform or What problems do we solve? 

A.     No dependency on platform , customer can shift between platforms as and when he needs and as per his economies at anytime. 

B.     Secure collaboration and access through browser to authenticated users for each element data, model & visualization. 

C.     Each model runs isolated in container , same instance can be sliced to provide multiple containers automatic way. 

D.     Proprietary data & model supported at initial set up. 

E.      Security and control

F.      Our main pitch the data scientist models, everything else by the platform irrespective of cloud. 

G.     Additional functionality such as data cleansing, data preparation as add ons.

5. How do we compete if AWS or Azure provides same service ? Our typical customer we believe is a 50 member team hedge fund  

A.     Data modeling is important but the infra provisioning needs hand holding .

B.     They use proprietary tools and data

C.     Support is the key, public cloud support is costly

D.     There is no need to stay wedded to a public cloud provider, should have the ability to switch based on economics.