Blog

Meet Eric Detterman: VP and Global Head of Products at LexPredict

At its inception, LexPredict consisted of Mike and Dan working closely together on strategies, projects, and consulting. Over the years, these initial efforts have bloomed into a full-fledged legal technology company with a whole suite of products and services.

As a growing company, LexPredict has thoughtfully expanded its team in order to seize market opportunities. To that end, we’ve hired a range of experienced and unique team members, and in a continuing effort to show how law and technology are becoming increasingly symbiotic, we will periodically profile members of our deep bench of highly-trained legal technologists.

This week, we’d like to introduce Eric Detterman, VP and Global Head of Products at LexPredict.

Early Work

Eric Detterman – VP and Global Head of Products at LexPredict

Eric began working at LexPredict about a year ago, but his working relationship with our CEO Michael Bommarito stretches back many years.

After earning a degree in economics, Eric worked in process improvement and process consulting. He helped large enterprises identify and implement best practices, working with teams both large and small to improve their efficiency. This people-driven focus brought Eric to Solid Logic Technology, where he initially worked with LexPredict CEO Mike Bommarito. Solid Logic’s wide array of clients included hedge funds, proprietary trading groups, and a large semiconductor company. This diverse array of companies all shared one thing in common: a commitment to data-driven processes. The mission of Eric and the others at Solid Logic: aid these data-driven companies in creating quantitative solutions, building predictive models and support for those models.

Eric got involved with one particular project centered around identifying best practices across a global sales team. Using gamification principles, Eric developed a weighted scorecard designed to incentivize best practices for salespeople and solution engineers.

Previously, the salespeople would go out and sell one device, when the company could actually supply 4, or 5, or 6. So what we worked on was … a structured way, on a web platform, for them to design the product and to identify – if you’re using X, you probably should use Y and Z, and to consolidate that and have a distributed view, similar to a Google Doc, where two people can interact with it simultaneously.

Gamification systems are often successful because of the innate human drives for satisfactory competition, skills mastery, and recognition. Through this experience, Eric learned how to apply gamification principles to enterprise software, and has brought that insight into ongoing projects at LexPredict that use those same concepts: LexSemble, and FantasySCOTUS in particular.

FinTech Innovation

Many current approaches to legal innovation got their start in the financial sector, where the usage of data-driven models and tools has a much longer history of successful implementation. With a degree in economics, specializing in econometrics, Eric was a natural at developing models and handling algorithmic trading. “I always had an interest in financial markets and quantitative development … everything I’ve done has had a predictive analytic or model-driven statistical slant, with a lot of infrastructure to lay in and fit models.”

For several years, Eric worked at various companies specializing in FinTech. He developed models, created investment strategies for ETFs and mutual funds, and developed software applications. One application he has worked on involved voice recognition software specialized for financial markets. “Bank A calls their broker who calls Bank B to execute a very large trade. But the financial terminology that a trader or broker would use on, say, a crude oil desk…to a layperson, they would not know what those terms are. Your common consumer-focused speech recognition tools, like Siri and Alexa, aren’t trained on that, either, so specialized tools are required.”

Plenty of industries rely heavily on jargon, from police dispatchers to hospital staff. Lawyers and other legal professionals also have their own field-specific dialects. Eric’s experience developing specialized speech recognition tools for FinTech, coupled with his experience building predictive models, positioned him perfectly for the machine learning and natural language processing efforts here at LexPredict (e.g. LexNLP, ContraxSuite, case prediction models, etc).

Joining LexPredict

Eric came to LexPredict in late 2017. He took the lead on designing and building ContraxSuite, applying his knowledge and experience with designing similar software tools in the past.

ContraxSuite and the LexNLP library were built and used as part of internal consulting projects, to solve a specific need, or do a specific analysis task. They were the culmination of a lot of different consulting projects.

What I’ve done since I’ve joined is to update a lot of the infrastructure components and to structure how they would be used in a larger-scale environment, in a cloud-first environment, while still supporting the legacy on-premises architectures of client companies.

– Eric Detterman

Eric has done a lot of the work on ContraxSuite’s backend components, while Kelly Marsh has led the expansion of the user interface so that anyone – not just a data scientist – can use ContraxSuite. The ultimate goal of ContraxSuite’s UI is that anyone who is comfortable with a word processing application can be equally comfortable working with ContraxSuite’s annotator and classifier training interfaces.

contraxsuite annotator

ContraxSuite Annotator

If you’re a multinational corporation and you wish to … identify agreements which would be subject to changing conditions such as Chinese tariffs or Brexit … that’s very difficult to get because the data’s not structured in an easily accessible way. But ContraxSuite, using the LexNLP library, can interrogate a set of documents for facts like … Is it Japanese yen? Is it South Korean won?  Is it US dollars? How do we extract that out? Once it’s extracted, we can use that data for model development purposes.

Open Source Development

When it comes to scaling ContraxSuite, Eric has focused on the underlying infrastructure and foundational components. While most can easily understand why a solid user interface is important for developing software tools, it requires the mind of a technologist like Eric to solve scaling problems in the background technical structure. Large audiences need to be able to use ContraxSuite without problems with lag or other errors. In addition, it’s been important all along to support both on-premise deployments of ContraxSuite, as well as cloud deployments. “We’ve consolidated that and begun using Docker containers for our deployment approach,” Eric says. Docker makes it so “we’re able to have a consistent architecture regardless of which infrastructure choices a firm makes, and so we can support their future growth going forward.”

Docker and GitHub host aspects of ContraxSuite’s framework. Methods for building intuitive user interfaces also come from the open source community. Deriving the full benefits of these open source tools is the main reason we decided, last summer, to open source ContraxSuite itself. The flexibility that an open source model provides is one of the things that attracted Eric to the project, and to LexPredict.

Leveraging powerful tools in the open source community, such as Docker, means that great developers can extend what’s been done without having to possess a specialized skill set. A coder, developer, or academic researcher – or even a hobbyist – can pull the Docker container, and have a fully-functioning system. An enterprise user can pull the containers in much the same way.

Looking Ahead

In the future, Eric intends to focus on spreading an understanding of AI technologies throughout the legal community. AI needs to be brought to the centers where it is most needed, because “many tools in the marketplace now assume there’s an easily accessible data set to identify items of interest.” This is in stark contrast to reality. An AI system can’t work in a vacuum; large, high-quality data sets are needed in order to train models.

To that end, it’s important to have “integration with document management systems, with contract management systems, with e-mail systems, with common fileshares … to make it much more easy for an end user to get the benefits of AI without having to leave their existing applications.” What ContraxSuite does well, it can do even better if it can be trained on larger data sets. And many firms want to have the benefits of AI – extraction, comparison, and other intelligence capabilities – and be able to integrate ContraxSuite with their own system of record (such as SharePoint, Google Drive, Office 365, DropBox, Box).  In other words, the future of ContraxSuite involves integration with other platforms. Eric is already working hard to make that happen.

 

About LexPredict

LexPredict is an enterprise legal technology and consulting firm. Our consulting teams specialize in legal analytics, legal data science and training, risk management, and legal data strategy consulting. We work with corporate legal departments and law firms to empower better organizational decision-making by improving processes, technology, and the ways people interact with both. We develop software and data tools, and also offer execution and education services.

LexPredict has a number of software and data products, including LexSemble, ContraxSuite, CounselTracker, and LexReserve. These products assist organizations with early case assessment and decision trees, contract analytics and workflows, outside counsel spend management, and case valuation. LexPredict also offers advisory and capital services for legal tech startups through its LexGen Ventures arm.

Contact LexPredict today to receive more information about our products and services.

Comments are closed, but trackbacks and pingbacks are open.