Many different types of people use Deephaven to develop data-driven applications and analytics. They tend to care more about progress and productivity than about whether their data happens to be an event stream, time series, transaction set, or batch file. They want their data engine to be high-performance, easy to use, and integrated with popular tools.
These are our people.
We particularly like it when our software enables them to quickly deliver results to each other or to piggy-back on one another’s work.
Deephaven is committed to serving as many data-driven use cases as possible from a single, empowering engine. To that end, the company holds the following principles dear:
- “Broadly useful and really quite good” beats “narrowly stunning”.
- Data is in a state of flux.
- Stream versus batch is a false dichotomy.
- Data and time are often two sides of a coin.
- Code must go to data, not vice versa.
- Single-threaded performance is important and requires constant investment.
- Modern systems scale.
- Copying data hurts.
- Teams need interoperable gear.
- The most popular roads should be paved.
- Ease of use and documentation matter a lot.
- Help when you can.
Long before the term “meme stock” was coined, Pete Goddard, Deephaven’s CEO, founded a quantitative trading company called Walleye Capital with a handful of partners. For many years Walleye focused on options market-making, an arms-race of an industry comprising teams of ML quants and low-level system developers.
In 2012, Walleye was interested in diversifying its business. While still focusing on a quantitative and systematic approach, it intended to grow the breadth of its trading strategies and move toward different frequencies of prediction. The partners knew the success of this new initiative would hinge on how quickly quants, data scientists, developers, and portfolio managers could discover and evolve strategies and signals. The business would win or lose based on its prowess in working with data.
With a few simple requirements in hand, the company surveyed the marketplace for a data system that would serve its needs.
- A single system that could handle both real-time and historical data.
- A single system that was powerful with both time series and traditional relational loads.
- A single system that was directly accessible (and helpful) to all of the company’s 100 employees.
- A single system that could drive applications and support analytics.
- A single system with a powerful back-end and a useful interface.
In 2012, nothing came close to satisfying those requirements. So we rolled our own.
In 2017, having witnessed the compelling story Deephaven powered at Walleye, Pete and six principal engineers spun themselves, the data system, and its related IP out of Walleye and formed Deephaven as an independent company.
Join our team
If you are a talented and highly productive developer, dev-rel, data scientist, dev-op, or cloud SRE; or a biz-dev, marketing or operations person that happens to be into software; then there is a spot for you at Deephaven. We hope you play well with others, write good unit tests and documentation, appreciate candid conversation, and get into the discussion about music on our Slack channel.
Deephaven today has three dozen engineers, a designer, a technical writer, and not much else. Though the company is headquartered in the suburbs of Minneapolis, the team is well represented in New York City and Colorado Springs, as well as family rooms, garages, and makeshift offices scattered around the U.S. and Canada.
We’re deeply committed to Deephaven’s mission, its customers’ success, and each other.