We’re kicking off our inaugural employee spotlight and who better to lead us off than one of Loris’ most tenured employees Seth Levine. Seth is our Lead Machine Learning Scientist and also hosts the podcast “Learning from Machine Learning”. We sat down with Seth to ask him ten questions – some random, some personal, and some professional – to get to know him better.
1. Can you tell me what do you do at Loris – but explain it to me in the most basic terms, so a 5-year old could understand?
SETH: We do so much at Loris! The simplest way I can put this is, I train machines to understand what is going on in conversations and then help people have better conversations now and in the future.
But if you’re interested in understanding a bit more – no matter what your age – Hillary Mason has a great video on this Computer Scientist Explains Machine Learning in 5 Levels of Difficulty | WIRED where she explains, “Machine Learning gives us the ability to learn things about the world from large amounts of data that as Human Beings we couldn’t possibly study or appreciate. Teach computers to learn patterns from looking at examples in data, such that they can recognize those patterns and apply them to new things they haven’t seen before.”
2. So, what are you most excited about right now, personally or professionally?
SETH: Personally, by far the most exciting thing right now is being a new Dad. Being with my daughter, who was born four months ago, and seeing a real neural network form in real-time is truly remarkable.
Professionally, the AI field is moving at a breakneck pace with new research, methodologies, libraries, and models being released daily. I am very excited about multi-modal models. These are models that have the ability to combine vision, audio, text, and other modalities as inputs and outputs.
Imagine a system that assists in creating immersive storytelling experiences. It could take a children’s storybook and not only read the text but also display related images, play music, and even generate animated characters that interact based on the story – all happening simultaneously, making the story come alive. Everybody learns and interacts with the world differently so having systems that can uniquely present information to individuals is very exciting. Multi-modal models bring us one step closer to making these types of use cases a reality.
3. Two-part question: Where are you from and what was your favorite (or least favorite) thing about your hometown?
SETH: I grew up in Great Neck, New York and the thing I loved about my hometown was some of the amazing teachers I had who really cared about their students. From an early age, I was given the opportunity to fall in love with learning. This helped shape who I am.
4. Keeping on the learning topic, what was the most interesting thing you’ve read or learned recently?
SETH: In an early chapter of The Alignment Problem by Brian Christian, he discusses how analyzing text embeddings over time can quantitatively show how human biases have changed over time. One of the most famous computational linguists Dan Jurafsky is on the paper: Word Embeddings Quantify 100 Years of Gender and Ethnic Stereotypes. By analyzing these embeddings trained on a century of text data alongside demographic information, it becomes possible to observe and understand shifts in social perceptions and associations. This might not be for everyone, but I find it very interesting.
5. If you had an entire day to yourself, how would you spend it?
SETH: Wow, that’s a good question. Outdoors and eating sushi. I would be on the beach. There is no better feeling than being in ocean, diving through waves. I would go to a park or a trail I’ve never been to. At some point, I would find myself sitting at a sushi bar and probably trying to strike up conversation with the sushi chef (usually to no avail). As long as there is good food and outdoors, I’m content.
6. What has been something you’ve purchased for under $100 that made a difference in your life?
SETH: A stand-alone guitar tuner so I don’t have to use an app on my phone. Then, I don’t get distracted by my phone when I want to play guitar.
7. Speaking of making a difference, why did you join Loris?
SETH: When I joined, Loris’ mission was to bring more empathy to conversations. I’ve always been interested in how humans interact, solving problems, and entrepreneurship. Loris was the perfect combination. After working in machine learning and as an Natural Language Processing (NLP) consultant, I reconnected with Etie (Loris’ CEO) and joined as a data scientist. Loris was an incredible opportunity for me to apply my Machine Learning background while wearing many different hats to influence all parts of the company.
8. For someone not in the field, what could they expect as a Data Scientist?
The main thing to expect is to be challenged with interesting problems. Our work is all about the intersection of human behavior and machine learning, so it’s not just about being a data scientist in a vacuum. You have to really understand human psychology – and empathy – to apply techniques that can ultimately help end users. Basically, you can’t train a machine to be empathetic if you aren’t already.
One of the things that I really love about Loris is that we have a really caring and smart team. Everybody takes pride in the work that they do. Every day I feel very fortunate to work with such a talented group of people.
9. Without disrupting the spacetime continuum, what advice would you give your 19-year old self?
My wife thinks my advice sounds like a fortune cookie, but here goes nothing. I would say:
- Trust your instincts – follow your intellectual curiosity
- Always keep learning, keep tinkering and trying new things
- You have to know the bitter to know the better
- Don’t sweat the small stuff (and most of it is small stuff). Ask yourself will this matter in a year from now or 5 years from now?
- Take a deep breath
- Life is a journey not a destination and the adventure is just beginning!
10. And for our final question, what’s one of your favorite moments working at Loris?
Working at a startup is the most amazing roller coaster. I’ve had so many amazing moments and spending time with the team is unbelievable. Having the chance to go to Israel and having the Israel team come to New York were very special opportunities.
One of the most memorable moments was my first day as a full-time employee, Etie surprised me with a mariachi band to welcome me to the team. I’ll never forget that.