Bringing Big Data Skills to Venture Labs (From Lab to Start-up World)
VentureLabs has a number of companies that crunch large amounts of data to create value. Data analysis and visualization is critical to many different business models these days.
Ibbaka Talent uses big data analytics to predict how well a person will fit a team’s needs (TeamMatch™) and to estimate how well a team as a whole matches the requirements and how likely it is to succeed (TeamFit™).
The precision agriculture company Semios is seeing its systems installed in about 20,000 acres this year. That lets it collect a lot of data. They want to use this data to develop better predictions and models for pest control.
Optigo Networks, who have a very innovative approach to securing intelligent buildings (intelligent buildings are one of the least secure parts of most company’s networks), use a lot of data from the buildings to detect and shut down threats.
Even Wearable Therapeutics (Snug Vest) is moving into data collection and analysis, and it will be an important part of VeloMetro’s operations as well.
So when two recent Ph.D.’s in physics (one from the University of British Columbia and the other from Brown in Rhode Island) showed up at VentureLabs we decided to offer them an office space and encouraged them to form a company that could level up the big data expertise at VentureLabs. I think this makes us the first incubator in Canada to invest in having two resident Ph.D.’s with big data expertise available!
We asked our two new guys about why they are coming into VentureLabs.
ML Infinity is Alex Loosley (from Brown) and Mike McDermott (UBC).
Why start your own business rather than join a company?
Starting our own business allows us to work on what truly excites and inspires us everyday. We both share a passion for data science and want the freedom to work autonomously and innovatively on challenging problems. From our discussions with several small businesses in Vancouver, we realized that many are in search of machine learning, statistical, and big data services, but are not yet in a position to take on a full time in-house data scientist. Starting our own consulting business gives us to opportunity to interface with these small and exciting businesses, grow our network, and to work on several inspiring technical projects over a short period of time. Furthermore, we have the opportunity here to learn about running a business. In the future, we feel that this experience will be invaluable.
Another reason to start a business is our strong friendship and proven ability to work together. We both have similar aspirations but took notably different paths to get to where we are today. The diversity in our knowledge and experience, put together, is an intangible that will help us build something special.
What did you learn from working in academia and studying that you think will help in the new project?
We have each worked at the frontier of knowledge creation, solving complex problems using techniques from a wide range of fields, from biology to economics to mathematics.
Having advanced research degrees not only helps us solve challenging problems, but also provides us a deep understanding of the problems themselves. We cannot only come up with innovative solutions; we clearly present how the solution works and why we chose it. As a result, the customer will be in a position to control, adapt, and scale the solutions we provide.
In the past, we have worked on large projects involving multiple institutions and many collaborators. We have learned how to communicate effectively with people that have a wide variety of backgrounds and needs. Finding common ground is key to any collaborative effort and we have learned how to do this. Problem solving with many professors over the years has also forced us to be agile. The same problem can have many solutions and some will better fit the needs of the client.
The academic process has taught us about the importance of maintaining a solid relationship with the people we are working with. Strong relationships foster a healthy environment and increased intellectual openness, factors that have a positive impact on the project. Academia got us used to working in a competitive environment where other groups were attempting to solve the same problem as us. When our teams were well managed and cohesive, we were able to succeed first.
Overall, working both autonomously and collaboratively to solve cross-disciplinary problems as part of a team was a central theme during our academic careers. Merging the best practices we developed to succeed in academia over five years with the need for a fast paced iterative approach in the startup world should allow us to find success with our new consulting business.
How can data analytics benefit Ibbaka Talent?
The questions that Ibbaka Talent are trying to answer are quite complex. The best implementation of an algorithm that simultaneously maximizes the potential for success of an ensemble of teams will require the integration of ideas from many different fields (sociology, economics, mathematics, communications, business, etc.).
The first step is to determine what the qualities of a successful team are based on, the components that make up the team. These qualities need to be measurable. Data analytics is the key to systematically identifying these qualities on a component-by-component basis through the integration and mining of massive amounts of data. Data analysis turns raw data into meaningful qualities.
Building on Ibbaka Talent’s existing models of what makes for good team fit, we can identify the additional sources of data that will have the biggest impact on prediction quality. The goal is to improve the model over each development cycle. Data analysis provides innovative techniques for integrating structured and unstructured data from both conventional and less conventional sources.
As an example, imagine an individual who has a reported skill (say project management) with 30 endorsements based on project experience. Today Ibbaka Talent’s SkillRank™ provides a ranking of skill level and confidence in that ranking (Ibbaka Talent being a Bayesian company). But is a number enough to give real insight into the skill. Imagine that this person has managed projects on Asana and Basecamp and that there is a large corpus of communication that could be analyzed. Natural Language Processing algorithms can be used to sort through all the text to more accurately score a person’s skill set and give a sense of their approach to project management.
And then, there is the issue of how teams fit with each other. The number of interactions between teams grows as the square of the number of teams and the efficient optimization of a large interacting system like this will require some big data techniques to run smoothly.
Data analysis techniques can be used so that Ibbaka Talent’s algorithms can improve themselves! In the pre-release stages one can have the algorithm learn how to make better teams based on simulations of team compositions and meshing, and post-release the algorithm will learn based on client feedback and other data (which can again be obtained by data mining techniques).
Wow. We are delighted to welcome Mike and Alex to the VentureLabs community. We think they will have a big impact.
