We build smart APIs using machine learning and natural language processing

People

Bradford Cross
CEO
Jason Wolfe
Research Engineer
Dave Golland
Research Engineer
Logan Linn
Backend Engineer
Conner Petzold
Product Engineer
Marco Munizaga
Product Engineer
Alyssa Carter
Backend Engineer
Patrick Redmond
Backend Engineer

Values

Take initiative

identify high value projects and make an impact

Problem solving

enjoy solving new problems more than making small gains on old problems

Transparency

communicate honestly and have empathy for yourself, teammates and customers

Play and have fun

we're all here to learn and grow together; don't take things too seriously

Join the team

Senior Backend Engineer

San Francisco

Are you a backend engineer with a passion for distributed systems and an interest in machine learning at scale?

Apply Now

Senior Research Engineer

San Francisco

Are you a machine learning researcher with a passion for distributed systems and engineering?

Apply Now

Research Engineering Manager

San Francisco

Are you a technical manager with startup experience and a passion for solving tricky data problems?

Apply Now

APIs

Our APIs are bundled as the Interest Graph.

Automatically identify the thematic content within a webpage using our topics API.

Explore the complex relationships between topics using our topic similarities API.

Accurately identify the structured content within a webpage using our aspects API

Recommend the latest, most relevant content from the web using our topic feeds API.

Strategy

Our APIs open up new possibilities at two levels: for developers and companies

For Developers

Our public APIs solve general problems like topic classification that enable building products and services that may not otherwise be possible within time, budget, and team constraints.

For Companies

We build custom solutions to problems using our enterprise APIs and our machine learning expertise. We solve domain-specific problems with direct revenue gains, such as predicting the value of a page impression for online advertising.

Current Machine Learning as a service landscape

The current approach to machine learning as a service that startups, Amazon and Microsoft have taken focuses on composable machine learning pipelines that you can configure and run in the cloud. These services provide lower level components such as regression models, optimization algorithms, and cross validation. The assumption is that these tools will empower teams to use machine learning to solve new problems faster.

What's different about our approach

We think the approach taken in the current landscape overlooks a lot of complications in building effective machine learning solutions. To deploy machine learning in production, there are many steps that simply cannot be efficiently automated in a way that scales to everyone's problems. These steps include data normalization, feature engineering, and model selection and training.

We're differentiated from this approach because we provide APIs with direct solutions to domain-specific problems rather than more general, lower level components.

At Prismatic, we've spent four years working on content recommendation and personalization. In that time, we've developed a number of internal machine learning and natural language processing APIs. We're releasing these APIs publicly, which we're packaging together and calling the Interest Graph.

For example, we have APIs that classify topics, recommend content, or learn custom models based on raw HTML input. This is an exciting business opportunity because we can work on solving interesting problems across many domains. These solutions give more leverage to a broader set of teams, and are therefore more valuable.

What's next

We build vertically-oriented enterprise products and services on top of our APIs. The first two verticals we operate in are online monetization and finance. These two verticals are a great place to start, because the revenue opportunities are clear and significant.

Our strategy is to grow revenue in these verticals quickly, and reinvest cash from these operations into deeper technology in the public APIs and broadening to new industry verticals. We call this our hub-and-spoke model, where cash flows from these spokes in specific industry verticals allow us to reinvest in exploring new spokes and deepening the technology in the hub.

Investors

Accel Partners
Venture and Growth Equity Firm

Accel Partners has been a top global venture capital fund since 1983, notably investing in the Facebook Series A, as well as enterprise business, like Cloudera, and SaaS businesses, like Dropbox.

Jim Breyer
Breyer Capital

Jim Breyer is the Founder/CEO of Breyer Capital, and a Partner at Accel Partners, who has been investing in Internet, media, and technology companies for over 30 years. He has a handful of 100X and many 25X investments including leading Accel's investment in Facebook in 2005, earning him #1 on the Forbes Midas List in 2011, 2012, and 2013. He sat on the boards of Wal-Mart, Dell, and 21st Century Fox, among others.

Yuri Milner
Mail.ru Group and DST Global

Yuri Milner founded investment firms Digital Sky Technologies (DST), now called Mail.ru Group and DST Global. Through DST Global, Milner is an investor in Facebook, Snapchat, Twitter, Zynga, Flipkart, Spotify, Groupon, JD.com, Planet Labs, Xiaomi, OlaCabs, Alibaba and many others.

Join the team

Senior Backend Engineer

San Francisco

Are you a backend engineer with a passion for distributed systems and an interest in machine learning at scale?

Apply Now

Senior Research Engineer

San Francisco

Are you a machine learning researcher with a passion for distributed systems and engineering?

Apply Now

Research Engineering Manager

San Francisco

Are you a technical manager with startup experience and a passion for solving tricky data problems?

Apply Now

Engineering Culture

Fascination with data problems from machine learning to distributed systems.

Know the state-of-the-art, but prefer simple solutions when possible.

Passion for functional programming and immutability (bonus points for Clojure)

Value engineering quality practices like design, automated testing, and production performance.

Participate in community with open-source, talks, blogging, and events.

Open Source

When we build something we think is awesome, we give back to the community with our open source projects. Our libraries for declarative programming, data description and validation, and reactive programming are among the most-used in the Clojure community.

Clojure(Script) library for declarative data description and validation.

A ClojureScript DOM manipulation and event library

A ClojureScript library of utilities for Om applications

Prismatic's Clojure(Script) utility belt

fnhouse is a library for writing Ring web handlers in a declarative, testable, and extensible manner

Blogging

We're active on our blog on the topics of Clojure, machine learning, and functional programming

We also post the results of our internal engineering practice meetings on GitHub.

Events

Deep Learning

Happy Hour with Sergey Karayev on Deep Learning and Computer Vision

Read the blog post
Clojure Community Night

Encouraging the Clojure Community to Connect and Improve

Read the blog post
Functional Frontend Frontier

or, How We Lost Grandma to Dysentery but Gained Stateless UI

Read the blog post

Join the team

Senior Backend Engineer

San Francisco

Are you a backend engineer with a passion for distributed systems and an interest in machine learning at scale?

Apply Now

Senior Research Engineer

San Francisco

Are you a machine learning researcher with a passion for distributed systems and engineering?

Apply Now

Research Engineering Manager

San Francisco

Are you a technical manager with startup experience and a passion for solving tricky data problems?

Apply Now