Atomic Intelligence is a boutique technology firm focused on text analytics and natural language processing. Our best-in-class hosted service automatically analyzes large volumes of content -- such as news, financial commentary, and specialized documents -- and detects topics, disambiguates names, quantifies sentiment, and extracts new information.

The Atomic service is used by investment firms who use the data as part of their proprietary trading strategies, by consumer-facing companies who need to understand what their customers are saying, and by technology companies who want to outsource text analysis.

Atomic Intelligence was founded in 2003 and is based in downtown San Francisco.

Atomic Intelligence's technology includes components for company name and product name detection, sentiment analysis, machine learning-based classification, term discovery, topic and theme analysis, and efficient keyword scanning.

We place equal emphasis on a flexible web-based user interface. Our unique web-based UI allows our customers and ourselves to see how extracted data relates to the original content, provide feedback so the system can learn, and configure processing. It is designed to be quickly adapted and extended for custom projects.

Our entire platform is written in Python. We make use of elegant, lightweight open-source code components such as MongoDB, SQLite, and CherryPy.

We typically provide access to our service as part of a consulting engagement which might include one or more of the following: project feasibility and specification, batch or live content processing, data analysis, algorithm development, and service set-up. Software can also be deployed on customers' in-house systems.

If you are interested in using the service in a product-only mode, without consulting, please do get in touch.

Atomic Intelligence consists of a small team of experienced software engineers with expertise in:

• Natural language processing, including statistical modelling, machine learning, and information extraction.
• Text processing algorithms, including tokenization, text retrieval algorithms, efficient text scanning, and character-based techniques.
• Prediction using linear and non-linear machine learning techniques
• Ontology development
• UI design
• Systems engineering, including hardware configurations for large-scale content analysis.
• Product management

Nicholas Haddock has a strong background in text and language processing, artificial intelligence, and user interface design. This expertise, combined with a keen eye for user needs, has produced a series of product and technology innovations.

Following a Ph.D. in Natural Language Processing at the University of Edinburgh, Haddock joined HP Labs Europe, where he worked on multimodal user interfaces for mobile devices.

As part of a collaboration with Karen Sparck-Jones, the inventor of Inverse Document Frequency, he developed a system for incremental free-text search which was similar in spirit to Google's Instant Search. (At the time Sparck-Jones said of free-text search, "No one seems to be interested in this stuff, I don't know why." Fast forward to the late 90s -- the situation had changed dramatically.)

This work led to a further prototype for a Personalized News Reader, which prioritized news items according to news-reading behaviour, and two systems combining speech with mobile user interfaces: Personal Message Manager, and Mobile Voice Notes.

Nick then led strategic planning for Hewlett-Packard's Home Products Division, with responsibility for future consumer products beyond the Pavilion PC. He went on to co-found HP's Consumer Internet Services Division, rolling out a broad suite of new product features and service partnerships, resulting in significant revenue growth.

View Nick Haddock's profile