FAQs
Technology
Our platform runs an isolated air-gapped configuration that can never see your production data. Instead model blueprints a deployed and inflated in a customer's production environment enabling all data to remain within the company's policies and systems.
Implementation timelines can vary based on project complexity. Typically, our implementation takes anywhere from 3-6 weeks for a base infrastructure deployment working closely with your teams, and we aim for efficient deployment within agreed-upon timelines.
Models are fine-tuned in parallel. Deployment strategies can be customized depending on customers' productionization practices.
We use a variety of techniques to reduce the exposure to hallucinations. Rather than one large model, we often leverage several fine tuned smaller models that run in sequence to provide traceable, auditable results.
Our model blueprints are trained and tuned on our own sourced datasets, and are deployed as blueprints which fine tune on your data.
This means that the models themselves are all uniquely customized for your specific requirements.
This approach also allows us to provide updated blueprints to support new techniques to provide a base case benchmark that continually improves.
No customer data is received or leveraged by Data Kinetic, meaning there is no leakage or risk of IP being shared between customers.
We do not maintain any customer data and only enable redundant and resilient protocols to support maintaining our own infrastructure.
Open Source models and libraries are run through security checks daily if employed, ensuring additional security from upstream attacks.
Customers
Only users with permissions originating from your environment can access the data. Data Kinetic does not see or have any customer data sent back to our infrastructure.
Generally most model deployments have an API to support faster integration. The core system and models operate with a private API framework that can be exposed on request.
Post-deployment adjustments are accommodated through our scalable and adaptable framework, ensuring flexibility in refining models as needed.
Model development timelines vary based on complexity. However, we ensure efficient and timely development based on your specific requirements.
Our dedicated support team provides ongoing maintenance, updates, and necessary support for the deployed models.
Platform agnosticism is a core tenet of Data Kinetic's value proposition. Our models are designed to be versatile and adaptable, facilitating deployment in various environments to meet your specific needs.
Our support team operates during regular business hours (8:00 AM - 6:00 PM) and offers extended support for urgent matters, ensuring you receive timely assistance.
Partners
Yes, white-labeling options are available for partners seeking to customize and brand our platform for their clients.
Level 2 and 3 Support for white-labeled platforms is provided by our team, ensuring seamless assistance to your clients.
You can share potential partnerships or interested contacts via info@datakinetic.com
Data Scientists & Data Analysts
We focus on empowering data analysts and data scientists with a comprehensive suite of tools for robust data analysis, models, and APIs to build additional use cases and customizations on top of.
Our platform is versatile and can seamlessly integrate with a variety of data sources, including structured, unstructured, and real-time data streams.
Data Kinetic does not see or manage any customer data, so security is handled within the customers own security context.
We tailor the onboarding process to suit your pace. Most customers are set up and operating the same day as signup.