Data Management & Analytics
Precise supports all aspects of data management, analytics, artificial intelligence and machine learning utilizing innovative processes, procedures, and technologies that help our customers to drive for innovation, operate with increased agility and improved performance. Our advanced data lake, data warehouse and analytics solutions deliver intelligence and actionable insights from structured and unstructured data. Our award-winning machine learning innovations promote better healthcare and food safety solutions.
AT A GLANCE
Large scale transactional database design
Cloud-based enterprise data lake, data warehouse implementation
Database migration strategy and implementation
Master data and metadata system design
BI, advanced analytics and machine learning modeling and development
Success Stories
Cloud Data Lake, Data Warehouse and Machine Learning Platform
Our customer, a large center of a federal agency, needed to establish a modern data lake and data warehouse to ingest and store large volume of structured and unstructured data from the agency, and its government and industry partners. It also needed a data analytics and machine learning platform to allow its business users and data scientists to consume the data, create BI dashboards and reports, and apply advanced data analytics and machine learning for deep insight via a self-service manner.
Precise team built the end-to-end platform that included data ingestion, cleansing, transformation, data lake, data warehouse, data catalog and metadata management, machine learning pipeline, self-service BI on AWS Cloud. We leveraged AWS and 3rd party service offerings such as Apache NiFi, Data Migration Service (DMS), S3, RedShift, Glue, SageMaker, Tableau and ThoughtSpot. The platform has successfully released to production within one year and the team is continuously enhancing the platform capability for new business cases.
Machine Learning as a Service
As part of an effort to accelerate our customer’s adoption of artificial intelligence (AI) and machine learning (ML), Precise engineers and data scientists built an enterprise Machine Learning as a Service (MLaaS) platform for a large federal government agency. The platform assembles a collection of pre-built models and algorithms addressing many of the agency’s most valuable use cases using computer vision, image classification, document understanding, natural language processing, and other machine learning technologies. The platform is hosted in the agency’s AWS cloud and leverages Google and Microsoft AI/ML services. The ML functions are published as Application Programming Interface (API) services to the agency’s entire user community. The platform has successfully obtained the agency’s Authorization to Operate (ATO) for High Impact systems and has been in production since 2021. The Precise team also built an enterprise AI/ML knowledge portal to promote AI/ML services, standards, best practices, and training among the agency’s user community.
Food Safety Outbreak Signal Detection via Social Media Analytics
In recent years, Precise has focused our R&D on applying machine learning to solve challenges facing healthcare and food safety. The Food Safety Outbreak Signal Detection system is one of the innovations from the effort. It has the capability to identify potential food outbreak signals from social media analysis. We developed the core signal detection portion using Nature Language Processing (NLP) Model BERT (Bidirectional Encoder Representations from Transformers) by Google. The system collects live updates from social media, predicts whether the text contains food safety concerns and extracts food type, sick symptom, and location if the text is relevant. This system also provides a Google Map interface that presents food outbreak heatmap on the map. Food safety regulators and inspectors can zoom-in to the exact map location that the outbreak signal is detected for details. This innovation won 2019 Government Innovation Award.
Counterfeit Drug Detector
Counterfeit drug detector aims to help consumers, pharmacists and regulatory agencies detect counterfeit drugs in the market to ensure public health. The solution we developed deploys an all-spectrum high definition camera to capture pictures under different lights with embedded Google’s TensorFlow Process Unit (TPU) into a processing board. We also built a convolutional neural network model (CNN) to analyze drug pictures locally. With this, we have achieved near real-time performance, and removed the dependency on network connectivity, making our solution viable in remote areas without network access. This patent-pending innovation won the Incubator Award and made to the top 8 innovation among 140+ applications at 2020 ACT-IAC Igniting Innovation Award.