Learn More About Hyperia Engineering and Machine Learning
Hyperia's ML team draws from IBM Watson and AlchemyAPI, and is responsible for building a multitude of scalable unstructured data analysis technologies.
Large-vocabulary speech recognition, speaker identification, diarization stack, trained on 20,000+ hours of speech.
Natural Language Processing
Topic extraction, sentiment analysis, dialog labeling, and text summarization designed especially for conversations.
Video analysis, optical character recognition, face localization, screen fingerprinting, and other visual recognition technologies
Semantic Search & QA
Multimodal (speech, language, image) vector search and semantic matching employing the latest in neural transformer models.
Hyperia's infrastructure team comes from AWS and Dish Network, and is responsible for our scalable & secure service backend.
Ingestion and analysis pipelines for Video, Audio, and Text. We build reliable and scalable services that serve millions of ML predictions per day using GPU-heavy workloads.
OAuth2, ETL, Webhooks, Websockets. We build robust integrations with third party apps and SaaS platforms to ingest and analyze business and commumications data.
Microservices galore. Queuing systems, RESTful services. We build, maintain, and deploy dozens of containerized services across our testing and production systems.
Configuration as Code
Continuous integration and deploy. We rely on automation for builds, deploys, logging, auditing, and more to manage our growing systems.
Hyperia's frontend development team has deep experience in streaming data analytics and human-computer interaction, coming from Amazon, Tastyworks and Tufts University..
We take the broad, complex pool of data available to our users and break it down so it’s easy to explore, and even easier to understand.
We make reusable, flexible patterns with room for user customization to provide a consistent and comfortable experience.