Look out Will Shortz, there is a new type of puzzle master on the scene. Although, the puzzle master I interviewed is not developing crosswords for the New York Times or to confound even the most astute puzzle aficionados on NPR’s Sunday Puzzle. Hanzo’s Data Scientist, Aidan Randle - Conde is an innovative problem solver who uses data from the ultimate puzzle, the world wide web, and pieces it together to uncover the truth and provide insights critical for legal decision making.
What do you do at Hanzo?
Aidan: I’m the lead data scientist at Hanzo and I oversee the data science development at the company. This ranges from improving the efficiency of our existing technology, developing new tools to improve the quality of existing and new products, and integrating machine learning into our products.
What problem are you currently solving?
Aidan: I am helping to integrate new tools into our Investigations product, which will give better results to our customers. Investigations helps clients who need to find and capture relevant information from the internet. It might seem like an intractable problem, but Hanzo has developed the tools needed to find the correct social media profiles and news articles for a given subject. Even then, the relevant content might be hidden within a vast volume of unstructured data across multiple platforms, so there are multiple tools to help zoom in on the most relevant content to help our clients find the answers they need.
What sets Investigations apart from other problems is that each step can act as a platform for another, even more powerful step. Finding the correct social media profiles for a given subject is the first step, and then using keyword filtering is the next, then we can develop models for different use cases, and build a very rich picture of the matter being studied. There is no limit to the complexity of reports to match the complexity of the cases we face.
I am also developing tools to make our quality assurance more efficient, by developing diagnostic tools, which will also help us monitor changes in content over time.
What do you like most about working for a legal technology company?
Aidan: It’s great being able to set my own goals and choose how to solve the problems in front of me. Hanzo has access to a very rich dataset and there are always new opportunities to add value.
For example, we can look at over a decade of data captures to work out what could cause a capture to fail and need resubmitting. Solving those kinds of problems can lead to significant savings of time and resources.
- How often do people share their location on Facebook?
- How likely are people to connect with their colleagues on Twitter?
- Which newspapers are the first to break a story, and which have more in-depth coverage?
We can go back through our archives to answer these questions and get better insights about where to look in the future. We’re limited only by our ingenuity and the amount of data we can collect.
Enjoying what you're reading?
I love your perspective on goal setting and limitations. Can you share a bit about your secret to success and your philosophy about problem-solving?
Aidan: I try to never isolate a problem from everything else I’m doing. Every time I find a solution to a problem I think about how I can use it to solve other problems, and how to use it as a stepping stone to creating an even more powerful solution to an even bigger problem.
At Hanzo, our Service Delivery Team ensures that pages are captured correctly, which is a very manual and time consuming process. I developed a simple tool to perform checks based on previous results, and suddenly found I had a tool which could also find changes on a page on two different dates. This would allow us to find when new comments are added, or when images fail to load. At the same time, it surfaces a lot of detailed information that I can use to train models to identify patterns in captures. It’s very rare that making progress in one area won’t help make progress in another.
Progress begets progress—showing truly the power of momentum. When you aren’t engaged in data science, what is your favorite thing to do?
Aidan: I love problem-solving, whether that’s playing or designing games, finding geocaches, or creating puzzle hunts. Anything can be turned into an elegant puzzle if you’re creative enough, and we are constantly surrounded by history, culture, language, and words. At home I have a large format printer from a home business I used to run. These days I use it when I develop puzzles for the local Puzzle Hunt group. It’s one thing to set up a puzzle event at the local art gallery, it’s completely different when people are folding origami models from huge A1 pieces of paper. I’ve also helped design both tabletop games and browser based games, and at the moment I’m developing machine learning models to help beat a few classic board games. I have to keep my mind active or I get bored very easily, and then I need to find a new project!
"I have to keep my mind active or I get bored very easily, and then I need to find a new project!"
Wow, you really are the Puzzlemaster! How did you get started in your career?
Aidan: I was a particle physicist, first as a Ph.D. student, and then taking part in research full time. That career took me all over the world from London to San Francisco to Geneva to Brussels, with plenty of international trips along the way. I helped with the Higgs boson discovery at CERN, and for me “big data” means several petabytes. After several years in physics research, I decided it was time to move on to some new challenges and found a whole new set of exciting opportunities at Hanzo.
Intrigued by Aidan? We are too. He’s working on cutting edge technology that will empower legal and compliance teams more efficient investigations.
Care to learn more? It's easy to set up a meeting today!