Sigmoidal

Key facts about Sigmoidal

Sigmoidal is a team of consultants and developers who, in just three years of the company’s existence, have managed to gain market confidence and currently have projects for such clients as T-Mobile, DARPA, Allianz, Kroll, NVIDIA in their portfolio. Moreover, the Sigmoidal was fortunate enough to collaborate with such giants as NASA, Intel, and Microsoft. As for the development process itself, when creating Artificial Intelligence solutions for its customers, the Sigmoidal team focuses on Natural Languages Processes, Signal Processing, Machine Learning, and Cognitive Computing.

Services provided by Sigmoidal

1. Software development

The team provides end-to-end solutions for the most diverse wishes of customers who, at first glance, come up with unrealistic ideas and problems. So, to satisfy all requests, the guys from Sigmoidal provide integration services for the Artificial Intelligence decision engines into user-friendly apps and well-documented APIs. 

To ensure a fast, reliable, efficient, and iterative development process, the team builds its applications with an admixture of Artificial Intelligence based on the Python as the best programming language for AI and Machine Learning. The choice in Python was not random since the most popular language for AI and ML differs in a great library ecosystem, low entry barrier, flexibility, platform independence, good visualization options, and more.

1. DevOps

To create the most flawless product, the team is actively introducing the DevOps policy into its workflow, thus ensuring the active interaction of development specialists with IT specialists and the mutual integration of their work processes into each other to ensure product quality.

To this end, Sigmoidal specialists using Docker and Kubernetes to deploy AI-optimized systems. Such open-source technologies are used by Google and Dropbox, and in the reliable hands of professionals, they usually bring the client a result that exceeds the requirements. Besides, in-depth knowledge of the use of NVIDIA GPUs makes the launch process of machine learning systems 10 times faster than competitors.

1. Security

Sigmoidal guarantees the safety of customers' private data, therefore they have trusted both strictly confidential company data and government data. An appropriate level of safety is achieved not only through well-trained personnel but also through frequent internal security audits. Besides, an important factor of such a security level is that the company's infrastructure operates in separate virtual private clouds, and data is encrypted.

1. Data Tagging

Given the features of Machine Learning, which requires training datasets, as examples for learning, the Sigmoidal collects and tags large datasets team using proprietary software with the Active Learning technology. This makes the process of collecting and processing and marking data faster and more accurately.

1. Machine Learning

Providing an extensive range of services related to machine learning, the team sets itself the goal of taking care of the client's project from the stage of data preparation to scalable deployment and production. When creating a machine learning project, the Sigmoidal guys offer features like NLP, Computer Vision, and Predictive Analytics.

The company provides Neuro-Linguistic Programming services or, more simply, Text Processing, to read information and data from human-language text, that is, not only code, but also figures, names, relationships, entities, and more. Besides, it is worth noting that the NLP provided by the company ranks the text based on relevance to speed up reporting and analysis.

In addition to NLP, the company is also strong in providing  Computer Vision development services - an area of ​​artificial intelligence that trains computers to interpret and understand the visual world, for example, to distinguish and detect objects in images, to classify images by thematic groups or by the presence of functions and more. Computer Vision is great for projects that need video surveillance systems, information management systems, modeling systems for objects or the environment, augmented reality systems, etc.

Sigmoidal uses statistical methods, data mining methods, game theory to provides Predictive Analytics services to predict the future behavior of objects and subjects. The most relevant technology for the fields of actuarial calculations, financial services, insurance, telecommunications, retail, tourism, healthcare, and pharmaceuticals.

The process of cooperation between Sigmoidal and the client, regardless of the goal, consists of 4 stages:

- Strategy workshop. The company organizes a three-day seminar at every clients' office as each project begins with an acquaintance and a thorough study of the client, who needs an AI solution for his business.

- Data collection & exploration. Given that Machine Learning needs a large amount of data that is necessary for training models, the stage of search analysis is a vital part of the development process. Typically, information and data are provided by the client, but the company can also collect the necessary data using online sources.

- Model development. Before the developing process, the team uses its resources to create a machine learning core - a model. Thousands of experiments conducted in parallel to deduce the ideal formula with which the client will be able to predict future trends or understand the semantics of the text.

- Full-stack application development. The last stage consists of the process of integrating a previously created model with a REST API or front-end application. The team takes on the process of developing all the necessary functions for convenient access to the model while ensuring scalability and a high level of security.

Case Studies

Consulting company Sigmoidal supposed to provide an AI solution whose goal was to transform and organize an endless stream of disparate sources, types, and volumes of available information. Collaboration transferred effective work both from Sigmoidal and from the client-side - risk management company with world-known name Kroll. Thus, thanks to deep research experience and decades of fieldwork, Kroll was able to create a system that improved and facilitated the company's rigorous and comprehensive search processes.

Thanks to the Kroll Due Diligence and Compliance service package presented by the company, the client was able to optimize their ability to provide customers with extremely relevant information and meaningful context in the shortest possible time. Collaboration with Kroll and its team continues, to providing an up-to-date and comprehensive basic data set that would allow analysts to focus more on adding valuable context and ideas that are most important to customers.

The second project involving the use of Artificial Intelligence in the pharmaceutical field. The client felt the need to collect information about the experience of using drugs through the use of non-traditional sources, such as social networks. The client found out that people are more willing to share their experience on forums than through traditional channels, so during the collaboration, Sigmoidal had to create a solution that could quickly scan various data resources such as Twitter, Facebook, and forum posts.

Thanks to the solution created by the Sigmoidal team, the process of detecting and processing reviews about the use of the drugs includes 4 stages. The first step is the actual appearance of the drug on the market and the process of using the medicine by people. The solution created by the Sigmoidal team comes into play from the moment social information on adverse reactions on the new drugs appears. When scanning posts on Twitter, Facebook and forum posts, the artificial intelligence algorithms, using advanced methods of processing natural language, filter out unnecessary information and determine data related to the topic to reviews about a particular medicine. At the last stage, the automatic generation of reports takes place. The client uses the provided reports to improve their medicinal products in the future.

Thus, with the final AI product, the client can easily make more effective identification of new adverse reactions to drugs, increase the safety of medicines, reduce the cost of drug safety surveillance systems and more.

Let's pay attention to another specific, real-life example project during which the Sigmoidal team provided an AI solution capable of detecting changes in the company's management structure. Before turning to a company capable of creating a modern solution to the problem, the client hired a group of research analysts who had to process a large amount of data, including thousands of web articles, tweets and messages on social networks to detect the latest changes in the structure of companies.

The solution provided by the AI ​​team can optimize and systematize the process of collecting and processing information and data. So, the client no longer needs a team of specialists who manually look for the necessary data, since from now on this phase is fully optimized. The system independently implements a constant process of receiving information ready for use by analysts. The second stage involves filtering out unnecessary and not relevant information using modern text classification algorithms, after which the system itself extracts knowledge from the text, showing only the most valuable information for analysts based on specified criteria. After this, the process of additional data processing, filtering, and prioritization takes place. Also, ideas are being generated using the information collected to predict other changes in the market and stock movements. The last step involves creating a clear and understandable report using Natural Language Generation technology, which allows AI to generate human-sounding reports.

Even though the technology was created for a specific client, it is very affordable and, if necessary, can integrate into the corporate workflow without any difficulties. So, the system is capable of not only infrastructure using the Sigmoidal API based on microservice but also be used as a separate cloud application for importing and exporting data to popular formats such as Excel, CSV, or Word.