Hyperbloom helped a leader in rare diseases and genetic medicines migrate their high-performance computing (HPC) cluster and data platform from an on-premises environment to the cloud.
Neuroimaging and Genomics Data
Cloud computing and DevOps
Our data practitioners are actively contributing to various aspects of the platform with a focus on cloud computing and DevOps activities.
Aggregate Analysis Processing
COINSTAC removes large barriers to traditional data-centric collaboration approaches. It enables groups of users to run common analyses on their own machines over their own datasets with ease. The results of these analyses are synchronized to the cloud, and undergo aggregate analysis processes using all contributor data.
Decentralized pipelines allow for distributed, iterative, and feature-rich analyses to be run, opening new and exciting capabilities for collaborative computation. It also offers data anonymity through differential privacy algorithms, so members do not need to fear PHI traceback.
Voice Computing for Healthcare
Hyperbloom architected and developed multiple software applications, web services, and data pipelines for NeuroLex. We supported their mission to make voice computing accessible to everyone. The automated machine learning workflows we built collected data, cleaned data, trained models, and deployed models.
We remove noise or augment voice datasets to be indicative of real-world use.
We can extract state-of-the art feature embeddings for audio, text, and image data.
We build deep learning models from voice data..
We manually or automatically label datasets.
We find new ways to monetize your voice data.
Use our API to clean, featurize, and model voice data on servers.
High Perfomance Computing and Data Platform Migration
A genomics company was storing data locally on SAN disks and performing analytics via Jupyter Hub hosted on their over-utilized HPCs. Hyperbloom wrote and deployed a software agent to their HPC cluster to act as a file gateway, allowing them to mount a file storage system with low latency access to their newly created S3 buckets. Hyperbloom implemented a site-to-site VPN for data transfer to save costs. We used a CloudWatch event to fire AWS Glue to transform the data into a columnar format, Parquet, and optimized it for the type of queries they planned to execute.