INTERVIEW: Armando Gonzalez, RavenPack's expert in applied big data & AI technologies
Armando is an expert in applied big data and artificial intelligence technologies. He has designed systems that turn unstructured content into structured data, primarily for financial trading applications. Armando is widely regarded as one of the most knowledgeable authorities on automated text and sentiment analysis. Here he answers our questions on the prospects for the alt data market and RavenPack’s latest innovations.
Can you share some examples of new ways customers are using your platform?
RavenPack is now being courted by fundamental and discretionary investment firms looking to get an edge from alternative sources of data. They use our latest tools to search across news, social media, and other textual content to generate insights for investing and trading stocks, risk management and compliance. Analysts are using our tools for comparative keyword research and to discover event-triggered spikes across billions of articles and filings. Combined with RavenPack’s sentiment scoring and analytics generated through Natural Language Processing (NLP), users can measure interest, in particular, topics across thousands of sources from around the globe and right down to the sentence level.
For example, our clients can easily determine how many times people mention a topic over a certain period of time to identify trends. They can learn which companies or products are mentioned with given keywords or what topics might be driving market interest. For example, we have clients very interested in tracking how their large cap portfolio investments are involved with blockchain technology. With our platform, they simply type the keyword “blockchain” and “US Large Caps” and get a dashboard filled with powerful analytics.
There has been a lot of talk about NLP and how advances in technique can open up access to vast information resources. What are the most exciting and innovative ways that RavenPack is incorporating these strategies into its offering?
RavenPack is a powerful tool for financial professionals because it can allow them to explore the magnitude of different events and how markets might react to them - all on demand. Investors and traders can quickly examine interest in a particular topic over time, where it’s most mentioned, with which stocks, and the related issues that are being discussed. Knowing what people are talking about provides a unique perspective on what they are currently interested in and influenced by.
The RavenPack Natural Language Processing (NLP) engine is the most sophisticated system currently in use by financial institutions. Within milliseconds, it can derive analytics data from tens of thousands of sources of information, including news and social media, and dig deep into electronic archives that span decades. Analytics are produced on an entity basis (by company, product, commodity, etc.) and include scores for relevance, novelty, and sentiment. The database also includes granular topic tags, relationships, and temporal analytics.
Data Scientists (Quants) can leverage RavenPack’s NLP technology to systematically read news and access the analysis as machine-readable data. Research analysts can easily create indicators of business, economic, or geopolitical trends in seconds. Finally, Portfolio Managers (PMs) can quantitatively factor the effects of news and public information into their fundamental investment and risk models.
Top tips: How can a Head of Quant or Head of ML best educate their CIO/CEO and Board to maximise budgetary sign off and input?
The best way for a Head of Quant or Machine Learning to educate and convince their CIO or CEO to buy a new dataset is to “show them proof.” The proof comes from proper backtests over long periods of time, overcoming business and economic cycles, bust and booms, and periods of uncertainty. They must demonstrate that the data sets are point-in-time sensitive and do not suffer from forward looking biases.
It is also important that the historical data was collected with proven integrity and there wasn’t any data backfilling or gaps that could influence the results. Also, they need to work with companies that have a proven track record to deliver “live” the same quality and performance seen historically. More importantly, any new purchases of data must demonstrate their “orthogonal” or additive value beyond what the firm already has or captures with existing data and tools.
What can be done about the talent war in AI and machine learning and how do you handle this in your organisation?
Talented Developers and Data Scientists are looking for the next big challenge. They want to join an organization that has innovation in their DNA. They want to be part of a team with a vision that has social meaning or has the opportunity of disrupting their particular field. RavenPack has been extremely successful at attracting AI talent not because it’s a large tech company, located in Silicon Valley, or because it offers huge salaries and bonus structures like those paid on Wall Street. The company has succeeded because it really cares about the people it hires, it entitles each person to own problems no one else has solved yet, and encourages them to be first at solving them. By settling in a comfortable and relaxed environment like the Spanish Mediterranean coast, developers and data scientists are more relaxed, in many cases bringing over their families, with a setting offering them more time to think and be creative - than if engulfed in the daily stress of living in a metropolitan area, like where you will find most tech companies these days.
What are you most looking forward to at the AI & Data Science in Trading conference?
I want to learn more about the challenges faced by the buy-side in using AI and how advanced they really are at successfully incorporating these new technologies into their workflow. In particular, I would like to know which types of data sets they find add the most value, where and why. I’m also interested in gaining perspectives on the future of AI and how experts see data science evolving in finance over the next decade.