deep learning in the wild
Deep Learning in the Wild

Infinia MLMachine Learning Solutions for Business

Who are you and what’s your role in the company?

As the Chief Scientist at Infinia ML, I oversee all of our company’s technical operations and research. I’m also the Vice Provost for Research at Duke University, where I’ve spent over two decades as a Professor of Electrical and Computer Engineering. My research has been widely published, and I was the most-published AI researcher at NIPS 2017.

I earned my BS, MS, and Ph.D. in electrical engineering from the University of Maryland.

How did you get in touch with deep or machine learning in general?

My initial interest was in electromagnetics, but in my early 30s I realized most of the field’s major innovation had already happened. So I did something unusual – I pivoted to where I could have greater impact which, for me, was machine learning.

Describe what your project/business does. Which problem are you solving?

We are a team of machine learning experts focused on delivering business impact. We help companies solve some of their toughest data challenges and unlock their biggest data opportunities. Many teams don’t have the expertise or the bandwidth to tackle these challenges – in fact, they may not fully realize what’s possible.

We bring deep machine learning expertise to projects across a number of industries including healthcare, human resources, legal, and manufacturing. Our algorithms help companies:

  • Prevent security breaches, product defects, and medical errors by scanning images and videos for anomalies that humans might miss.
  • Extract meaning from large numbers of complex documents, automating a repetitive task in a fraction of the time.
  • Perform biomedical analysis to detect genetic risk patterns without invasive procedures.
  • Automatically align unique data into a standardized industry framework, saving manual effort and allowing more accurate analysis.
  • Visualize reality by modeling 3D images from 2D counterparts.
In every case, our goal is to help lower costs, increase efficiency, and/or achieve a new breakthrough. Many of us come from academia, but everything we do must have a tangible business impact.

Can you describe your tech stack, what libraries hardware / architecture are you using and why?

We look for prospective data scientists who have experience with Python, including libraries such as NumPy, SciPy, pandas, scikit-learn, matplotlib, TensorFlow, and NLTK. We work with deep learning models including CNN and RNN architectures.

We also look for experience working with large datasets, including NoSQL and relational databases, experience with Linux, and experience with cloud computing and Amazon EC2.

What were some of the biggest challenges technical or business related you encountered and how did you overcome them?

Getting the right client data is of course an essential and ongoing challenge. So is helping clients understand what is actually possible. They may not realize all the ways that their data could help them lower costs, increase efficiency, and achieve new breakthroughs for their internal operations, or external products and services.

Machine learning problems that appear superficially different may actually have similar solutions. So we might spend time showcasing a variety of different use cases in different industries just to help clients spark their own ideas.

How has the industry response been to your solution?

I was getting a lot of inbound requests at Duke to help companies solve machine learning and data challenges.  That’s why we started the company.

Demand for our services at Infinia ML is very strong. We are working with a variety of Fortune 500 companies on a number really exciting projects.

What excites you most about the future of deep learning? Where do you hope the technology is going / where do you see the greatest opportunities?

I’m excited to see our field, which has been around for decades, making a real, tangible impact in the world.

I compare its place in society to the mobile phone. In the 1990s, many people were aware of mobile phones, but not everybody had them or felt they needed them. Today, mobile phones are ubiquitous - they touch every industry and we all wonder how we ever lived without them. That’s how people are going to feel about machine learning/deep learning in 10 years.

How do you keep up with new papers and publications?

I’m reading (and writing) all the time! Advising the up-and-coming researchers in my machine learning lab at Duke is a key way I keep on top of the latest developments.

What do you wish you had known before getting into deep learning?

I wish I’d understood just how meaningful and interesting the work would be – I would’ve gotten into it sooner!

What do you wish to see on a candidate’s resume who is applying for an engineering position involving deep learning?

Beyond the obvious technical proficiency, we’re looking for people who can thrive in a fast-moving startup environment. Unlike a large company, team members may work on multiple projects at once and may be asked to pitch in on other projects that are going on. We work in a number of industries, everything from healthcare to legal to marketing to people analytics, so we appreciate candidates that want that variety.

To find the right mentality, we have a preference for candidates who have successfully worked at fast-moving companies before.

What do you recommend candidates do to make themselves more attractive to employers looking to hire a deep learning professional?

At Infinia ML, we’re focused on the practical application of academic research. The more you can show us that you are too, the more employable you’ll appear. That might mean having done internships or even interesting side projects that show your ability to put theory into practice.

Can you give us some facts about your current team and point out, why a candidate may want to join you?

We’re growing rapidly and are working with some of the largest companies and organizations in the world. This means that new team members can have a massive impact not just in our company, but in our industry.