What we do

Integration is the key for a successful data platform. Systems are becoming increasingly complex and data management brings new challeges on size, speed and features of the technological components. Over time companies that adopted a tactical approach with point-to-point and ad-hoc solutions ended up on relying on spaghetti architectures that quickly accumulated a large technical debt. This debt usually doesn't allow the quick evolution needed by the current speed of business. For this reason in Quantyca we don't take shortcuts: we take the highest care while developing integration pipelines. To move data between systems we deploy Kafka as the backbone of the data platform and, whenever possible, we employ an event sourcing paradigm. To consolidate and transform data inside the various systems Talend is the technology of choice. On all realtime offloading scenarios, both for modern or legacy technologies like SAP or mainframe we leverage the high performance CDC technology of Attunity.

Big Data is not just Hadoop, but enable a polyglot architecture. Hadoop is well suited to store large quantities of structured or semi-structured data and process them in batch mode. It is much less suited to manage other workloads, like high frequency transactional with many updates or analytical tasks with low latency requirements. The only way to takle that is to embrace a polyglot architecture that employs other storage technologies, both RDBMS and NoSQL, along with Hadoop. We propose Cloudera as an Hadoop distribution, Snowflake and Vertica to manage the low latency requirements, Neo4j for graph analytics.

Migration to cloud of data and computation is a continuously increasing trend. However, not all applications will move on the cloud in the near future. Some legacy applications and systems like ERPs or mainframe do not have a clear ROI. That's the reason why hybrid cloud architectures are on the rise. Quantyca helps the customer in designing hybrid architectures that maximize agility, scalability and optimize the overall ROI. Google Cloud Platform is our main partner on this journey.

We believe that Data Platforms need to be integrated in the overall It infrastructure not only for data ingestion and human consumption via reporting/visualization, but also using services to enable machine to machine interactions. Data platforms must interact with other applications to distribute in the broadest way possible any kind of data, from the atomic transaction history of a customer to kpis and scorings generated by ML and DL algorithms.

Historically, data integration has been the domain of ETL batch processing and system integration the domain of ESB architectures. Nowadays the landscape is more blurred, given the rise of the microservices paradigm in application design. A succesful architecture, that often has to include legacy applications, has to be designed with a broad view with a clear balance of the different paradigms.
In Quantyca we have thorough competences on all the tecnologies and patterns, like Java/Spring/Kubernetes/Kafka for microservies, Talend ESB for a traditional SOA architecture and Tibco Mashery and Google Apigee for API gateway.

Machine Learning and Deep Learning are continuosly proving their impact in various functional scenarios. Data Science is increasingly expanding its role in many organizations. However, Data Science activities are too many times separated at an organizational or technological level, and challenges related to actual deploy in production are underestimated.

The Data Science practice of Quantyca has been developed starting from this evidence, leveraging these three success factors:

• A strong focus on Machine Learning and Deep Learning theory, with deep knowledge of the most productive technologies in the field (Fastai/Pytorch, Scikit Learn and Tibco Data Science)
• Ability to define with a pragmatic approach the Business Cases for the application of these methods, not only with a predictive mindset, but most importantly to optimize the key factors of each Business Case
• Competence and attention on the aspects related to the entire lifecycle of the model, from developement to deployment and maintenance, with a structured approach to model industrialization and serving

In increasingly complex and distributed IT environments between on prem and cloud it has become necessary a radical revision in the way of managing the infrastructure in order to be able to conceal often opposing needs such as scalability, high reliability, security and speed in the setups (provisioning and configuration) . At Quantyca we have developed infrastructure management, development and automation skills aimed at making the data platforms we develop more robust, faster in evolution and easier to maintain in production. Specifically we use Docker and Kubernetes to isolate the computation, moving it and scaling it when necessary. We use SaltStack to automate infrastructure provisioning and configuration. Finally, we use the ELK stack to monitor the status of the systems.

A data platform is not effective if it is not able to support decisions , generate actions and / or monetize information. The visualization of the data by means of reports, dashboards or free search tools supports the business in the extraction of value from data or, more generally, from the underlying platform. At Quantyca we have a consolidated experience in terms of data visualization . Specifically we use Tibco (Spotfire or Jasper Report), Microstrategy and PowerBI for reports and dashboards, Linkurious to visualize and navigate data storend in graph database like Neo4J.

The challenges in Data Management projects are not only of technical nature. Complexites are also organizational, human and regulatory. Creating a successful Data Platform therefore requires a strong governance of the data lifecycle to ensure agility in developement but also quality and reliability of the data. Regulatory prescriptions (i.e. GDPR, e-privacy and vertical market regulations) also require a deep control and knowledge of data and processes.

Our approach is based on a systematic application of best practices and through application of regulatory prescriptions in every component of the Data Platform, to increase constantly the delivery quality.

• Using metadata management solutions we define and document the semantic meaning of the data and clarify the lineage of the transformations. We leverage as a shared layer all data with a common and cross-functional meaning: master data management.
• We apply and document best practices on security, encryption at rest and access mangement.

As data and technology becomes more and more critical in the value generating processes, the ability to deploy custom data products and applications, along with commercial packages, becomes critical as a differentiating factor. Quantyca develops custom applications tailored to the customer needs using Spring Boot for the backend, React and Flux for the frontend, Kafka Streams, Scala and spark for analytics realtime data processing and Python for model developement.