Unscrambl OXYGEN™

Unscrambl OXYGEN is the infrastructure you need to start surfing the Big Data wave.

It is our scalable stream processing engine that allows the development and deployment of applications for data-driven businesses and industries.

The key goal of stream data processing is to refine incoming data and take action upon it, all in real-time. The top three elements of a stream processing platform are: Programmability, Performance and Pliancy.

Programmability refers to the support for rapid prototyping and application development, allowing you to turn ideas into code into action.

Performance is measured by throughput and latency, enabling you to go from handling one person’s data to a million people’s data with ease.

Pliancy is the ability of the platform to keep adapting, transforming and evolving itself and most importantly, to keep running in the face of ever-evolving data, requirements and environments.

Applications on OXYGEN continuously and simultaneously process a wide variety of structured and unstructured data sources and convert them into a canonical format. A variety of analytics is then applied on them and action is taken on the data in real-time. These applications can be written in a variety of languages, with prime support for Python, which ensures minimal turn-around time, to code for updates and future developments.

The platform also includes in-built support for fault-tolerance, and reliability, and the ability to incorporate new code modules and logics into running applications in a seamless fashion.

All this means persistent quality solutions and quicker actions at the time they matter most.

Unscrambl BRAIN™

Unscrambl BRAIN is the cognitive mastermind behind our platform. If OXYGEN captures data then BRAIN makes sense of it, giving it meaning and thus formulating recommendations based on the data and even acting upon what is recommended.

This real-time analytics engine is designed to see patterns, find anomalies, reflect on insights, generate ideas, remember, recall, juxtapose, contrast, predict and learn – all the time, in real-time.


Time-series aggregated store
Extremely space-efficient, based on a patent pending approach developed by our team. Recharges, Number of dropped calls, Number of international calls, …in the past 10 minutes, hour, day, week, month, year, lifetime.

In-Memory profile store
Fast read/write access to features that are associated with an entity, Last known location of customers, Predicted home and work locations…

Python-based ML framework for building and online scoring of models
Sepsis Prediction Model Anomaly Detection Models, Next Best Action

^ Back to top