Q: Can you explain exactly what your company does?
A: Infinia ML applies and audits machine learning. Our advanced ML experts are focused on the analysis of documents and large volumes of text (invoices, contracts, emails, medical records, customer support tickets, research papers, etc.). Leveraging our proprietary technology, we have automated much of the text processing pipelines, as well as context-based data extraction and classification. Unique to Infinia ML, and critically important to our machine learning approach, is our process of evaluating and auditing the performance of our models. We’ve built software that exposes key information (output metrics, logs, input schemas) of a machine learning pipeline, and we combine that with expert services to ensure that machine learning models are operating consistent with expectations. What our clients get are algorithms that transform their business, but also algorithms that perform over time with minimized risk of glitches or biases evolving in the machine learning pipeline.
Q: What insights do you bring to your role as CEO?
A: At Carrick we’ve studied machine learning and artificial intelligence from the outside in. Most of the market – investors, operators, consultants, analysts, etc. – are trying to fit AI and ML into a traditional software mold. But AI and ML are not “products” in a traditional sense. AI and ML encompass a series of approaches that use data to solve problems or make predictions. Consequently, machine learning models are not static. They are more like living, breathing organisms that constantly evolve based on the data they ingest – code that generates new code based on data. This understanding that Infinia ML cannot operate like a traditional software product company, and our appreciation of the responsibility to monitor and audit how our technology evolves, has shaped our culture and our strategy for building the company.
Q: What makes Infinia ML different than other machine learning companies?
A: What differentiates Infinia ML is that we are grounded in the reality that machine learning can never be fully automated, and we commit the resources to monitoring and auditing the process to ensure uncorrupted outcomes.
Many companies have raised money to build a general-use AI platform; where citizen data scientists or data analysts can leverage the platform to automatically build AI models that are effective across different use cases. The expectation is that these AI tools can be delivered to teams in the same way that Salesforce delivers CRM tools to sales teams. The other approach you see out there is to deliver pure AI services, where everything is built from scratch. This is the way data scientists typically prefer to work, with lots of trial and error to get to an optimal solution. Infinia ML operates as a hybrid of the general-use platform and services approaches. We have built a set of generalizable components for processing data and building ML applications (and a framework for deploying them) but we take those things and customize them to each use case. This leads to outcomes that are efficient and impactful – a rarity in AI/ML today.
Q: What role can machine learning play in the Covid 19 and post-Covid 19 environment?
A: Machine learning was already progressing at a rapid rate, but with increased digitization, there is a heightened opportunity for ML to support business processes. Recently, there has been an acceleration in the already massive amounts of data being generated by businesses. Many traditional workflows (e.g. Supply Chains) have been tested during the pandemic and the world is learning to operate intelligently based on data and more intelligent design of workflows.
At the same time, many industries are struggling to survive and are operating on razor thin margins. Machine learning helps organizations by creating efficiencies that reduce costs and increase revenue through better decision making.
Q: What are the benefits of working with a group like Carrick Capital Partners?
A: Carrick’s operational expertise is incredibly valuable, but what truly sets them apart is the agility and flexibility of their approach. They do not subscribe to the mindset that there is only one way to build a company. If you look at the best advisors in any discipline, the one constant is the ability to adjust to the strengths of your team. For example, the best basketball coaches in the world are not “system” coaches. They are coaches that take principles and then adapt/generate a scheme that fits their players. Like a good coach, Carrick gets to know every management team and identifies where they need help and what scheme will be most conducive to their success.
Lots of Carrick’s peers are more playbook oriented. Carrick offers expertise in finance, go-to-market, restructuring, and operations, but a given portfolio company may only need support in one or two of these areas. Because Carrick is driven primarily by delivering the best outcomes for their portfolio, they don’t need to put their footprint on everything. That is a big differentiator. A simple hack for discerning this about the Carrick team is by noticing how good they are (collectively) at listening and asking questions. It’s a team of listeners.
Q: What’s your advice for other businesses considering private equity?
A: If you have a good company there will be lots of interested PE firms, so it really comes down to who you trust, who you think you can work with, and who can help you outside of capital.
Carrick only invests when they know the business and market well, and when they can see how their expertise is suited to helping the company reach the next level. All the entrepreneurs and management teams working with Carrick chose them because Carrick knows their markets, understands their key challenges, and is well suited to help.
When entrepreneurs are raising money, they should ask themselves: “Does this investor uniquely understand what I do, and can they help me? What benefit can they provide?” I would also suggest they seek an investor who is willing to disagree with them. When I go to Carrick, I am always challenged to at least think through a different POV. They don’t tell me what I must do, but they don’t always tell me what I want to hear either.
Q: What are your goals for Infinia ML over the next five years in terms of both growth and measurable business impact?
A: We’ve already established a great foundation of clients and expertise. One of the things we set out to prove was that there is a lot of talent in the Research Triangle, and that you don’t need to be in Silicon Valley to work on the cutting edge of machine learning and artificial intelligence. The next phase for us is about maturing as a company and continuing to have a laser focus on applying our machine learning toolset to make business impact.
There’s all the obvious stuff we are building toward: client growth, sales growth, solution-market fit, proving our business model can deliver profitability. These things are all indicators of our larger goal, which is to become the premier vendor for document and text-based machine learning applications, as well as the trusted 3rd party for auditing machine learning applications.
Q: What are you most excited about at Infinia ML?
A: Everything in business and technology is a derivative of the people involved. First and foremost, I am really excited about the quality of our team. We have 30 data scientists and ML engineers who operate at a uniquely high level. It started with world class machine learning scientists Lawrence Carin, PH.D., Infinia ML’s Chief Scientist; Ya Xue, the VP of Data Science at Infinia ML; and the founding team. People in this field are drawn to others who can help them push the boundaries of the way they think. The research and work our team has done in the past helped attract more highly qualified technologists who would otherwise have been very difficult to recruit. By starting with exceptional people, we have been able to cultivate and nurture a unique team and culture. The caliber of our talent also enables us to do exciting work. We have cracked the code on what it takes to develop real-world machine learning applications, not just science projects. As a result, we have developed market transforming intellectual property and have a head start on solving some big problems.
