Deep learning model to bring insights and decision capabilities to OCPBilling - ON-544Project type: Research
Desired discipline(s): Computer science, Mathematical Sciences, Business, Social Sciences & Humanities
Project Length: 6 months to 1 year
Preferred start date: 01/01/2022
Language requirement: English
Location(s): Toronto, ON, Canada; Canada; Canada; Canada
No. of positions: 2
Desired education level: Master's
About the company:
EAIPure is a 2-years old startup with more than 15 employees. The company is a spinoff from EAIBrasil, a software integration company with over 20 years in the market and more than 100 employees with expertise in airports, telecoms, and financial clients, among others.
EAIPure’s key product is OPCBilling, created to seize a market opportunity left by outdated benchmarks platforms which handicap time to market offers in complex and dynamic billing environments.
OPCBilling gives the freedom to create almost real-time offerings without traditional IT bottlenecks and no coding required. In addition to not requiring IT resources for dynamic updates, OPCBilling supports EAIPure’s clients' IT departments by bringing visibility and control over the entire IT environment relating to Billing.
By removing internal administrative and IT hurdles, the platform allows clients to focus on revenue generation opportunities through the evolution of service offerings in fast and flexible manner.
Our next step is to incorporate AI capabilities into OPCBilling, to generate deeper insights into complex billing environments, identifying timely revenue opportunities in operations with fast-changing conditions with time-sensitive revenue opportunities.
Describe the project.:
The OPCBilling platform allows customer offers to be configured based on products, services, and information related to consumption and availability of the customer environment (SRE).
OPCBilling was created with clean structuring of data to be made available for the training of machine learning (ML) models.
A ML model will be created to be configured during new product creation. The model can be trained with abundant information that is already available from the structured data gathered by OPCBilling.
This new ML billing model will allow companies to improve their offer of profitable products/services in a fast-paced dynamic and flexible billing environment.
It will be a configurable machine learning model that provides insights to companies about their dynamic product offerings. The model must be able to adjust the configured offers and understand their characteristics related to sales, consumption, and availability from the unformatted data of OPCBilling. It must be able to bring insights to support improvements in offerings and confirm whether these improvements have actually taken effect. We want to give OPCBilling the necessary intelligence to support our customers in the deepest vision about their product/service and about characteristics that would go unnoticed by a human being without the insights provided by machine learning algorithms. We're going to transform the culture of companies and focus on what you really care about, which is to create the best possible offer.
Methodologies/techniques to be used:
- Data Science
- Machine Learning/deep learning
We are looking for a researcher who is able to:
- Understand the concept of OPCBilling, data model and suggest improvements in data layout to meet machine learning best practices.
- Understand the main insights we want to develop (indicated above), propose new ones, and classify as to the effort impact on the business.
- Point out the best solutions/concepts to be implemented.
- Develop Machine Learning models, configurable that meet all insights (remembering that it must be intuitive and configurable).
- Suggest the best platforms and practices to develop and test models.
- Deep knowledge in Machine Learning, deep learning, and data science.
- Being able to indicate the best platform to be used in the established model and tools available in the market, helping to design the tool's roadmap.