USF MSDS Minicast
A conversation where Daniel discusses his application based approach to data science and the vast multitude of applications including AI generated synthetic media, healthcare, and cryptocurrency.
Daniel Blessing joined the MSDS program after obtaining his bachelor’s degrees in Applied Mathematics and Statistics and Chemical and Bimolecular Engineering at John Hopkins University.
Can you start by telling us a little bit about your background?
Daniel: Absolutely. I am a 2020 graduate from Johns Hopkins with a degree in applied stats and chemical engineering. And I originally was interested in becoming a MD. But as I went to college, my passions more kind of went towards statistics and computer science. I felt like if I was on the fence of whether to become an MD or not, then it’s probably not the job for me. And so, once I graduated, I had a few different options to go into industry or continue my education. I looked at what the options were and data science and CS were the places to go that gave you the best options, the best jobs. I looked up some of the data science classes and they really interested me. So, I think that it was a great program for me because I’m really interested in statistics and how that can be applied. And then there’s a lot of opportunities to use data science in health care which would be a great way to kind of have an impact in the healthcare field but not be a doctor. So I applied to a few places and then I wound up at USFCA. So far, I’ve been really happy with the program.
Approach to Data Science
“Great. So how would you say you approach data science?”
Daniel: I generally approach it by thinking about what are the applications for it because I’m a very application focused person. I love thinking about different ways some of these new techniques could be used. So, I like to read newsletters about what’s going on in the field and how that’s being applied and then from there, I use my interest of, “oh I wonder how people generate fake voices” to motivate me to try to learn the the hard science. So I get the big picture of how it’s being used (where it could go) and then I use that to motivate me to go to these classes and read the text books and try to code in Python, some of these basic stuff. Does that make sense? Is there anything specifically, you want to know about how I approach it?
Christabelle: No that’s great yeah the deepfakes are crazy, nowadays. So hearing about the current state of the art and thinking about the implications are definitely very motivating, to understand how they’re generated. Very fascinating.
Daniel: Yeah and the deep fakes are just the tip of the iceberg. Obviously, that’s kind of the malicious implementation but there’s a lot of interesting things. If you can generate a voice or a video, you can have things like AI assistants where it’s companions for older people. Or just AI generated music where the songs are generated with a fake voice. I think OpenAI released a paper. I think it was March, around there. They generated some light pop songs, completely all AI generated, and it sounded really good. So yeah, those are a few of the things that I’m interested in, for the application side.
Why did you choose USF MSDS?
“That’s really interesting. Yeah that’s definitely a place where there’s going to be a lot of innovation in the future. So why did you choose USF MSDS? Is it primarily because it’s an application focused program?”
Daniel: Yeah, exactly that. There was a few reasons. I really enjoyed the course list and it was really application focused and only a year. Some other programs — I won’t name names — are less application focused and a little bit longer. And then, I really enjoyed the faculty. So I reached out to a few people beforehand and they were all super nice. Also, I really loved the thought of going to San Francisco and being in that world but obviously, Covid happened so I can’t really. That was an unforeseen circumstance but basically the course, the faculty, and the location. Oh — and a big part for me too was the job outcomes. USF is very clear about, “this is where people go and this is how much they make.” And I look at some other top-rated programs and they didn’t give statistics about the average salary upon graduation. Then, I looked at LinkedIn and I saw that a lot of graduates from top-level programs and USF work in the exact same job, except the top-level program costs way more and took way longer. And so, USF is a top-rated program I don’t want to say it’s not but just for more traditional top-level schools. But USF is a top-level program and if you look at the average GRE quant score, it’s very competitive. So, I think it’s a great program for people really interested in pursuing a career in data science.
Favorite Class and Project?
“So what has been your favorite class and or project so far and why?”
Daniel: My favorite class has been data acquisition, obviously. You know, it’s a class where you really got to build stuff. You got to build servers, different projects. I think my favorite project has been the twitter sentiment because it’s really interesting. It’s really topical. It’s something that you know, is actually useful, too. It’s not just a theoretical problem. It’s like, “let’s build a project of this twitter website where you type in somebody’s name and see their tweets and rank them using sentiment.” And something like that is actually useful in the real world. For example, a hedge fund might, each day, analyze the sentiment of all the tweets with the hashtag of a stock they’re interested in and see what the general retail sentiment is and use that to make decisions. So it was really interesting to work on a project that is something that you can actually do for a job.
What Projects Are You Most Passionate About?
“What kind of projects are you most passionate about or what domain do you see yourself working in?”
Daniel: Well, I do sometime want to work in the healthcare domain. But honestly, my focus right now is finance or cryptocurrency. I think they’re the most interesting things to work on right now. So, I could see myself working at a hedge fund doing some data science work for them. Also, I think a really interesting field that’s coming up is creative generation with AI, like we talked about. So, like the new GPT-3. Or is it 4? I can’t remember what the newest one is but some of the texts that I can generate is very amazing. So there are all sorts of applications for that, that startups and established companies are using. So, I would say I’ll probably end up in either three areas: healthcare, cryptocurrency finance stuff, or creative generation.
Where were you placed for practicum?
“So stemming off of that, undoubtedly one of the core aspects of this program is the practicum. So this is a one-year intensive program and a big reason people join is for the practicum experience where we get nine months of training in a real-world setting in a position that we might not have been able to get right out of the gate otherwise. Without giving too much away, where were you placed for practicum and what type of work are they doing?”
Daniel: So I got placed in the Stanford Medical School and it’s a data science practicum so I’ll be using cutting-edge techniques in data science to help solve a relevant medical problem.
Victor: Right, yeah. And as it turns out, Daniel and I are actually working on the same practicum at Stanford Medical. We were coincidentally partnered on a relational database project and found out we both chose this project and both got a sort of half confirmation back so we were definitely very confused in the moment but it all worked out, thankfully.
What do you do on your leisurely time?
“So, addressing the whole online thing due to Covid, you mentioned that a reason you joined USF MSDS was that it was in San Francisco. I think, in general, lot of people really wanted to experience university in San Francisco, especially if they’re not from the bay area. So given the unfortunate circumstances what do you do in a leisurely way to keep yourself sane during these times?”
Daniel: Yeah, that’s a great question. Well, I think a big thing is exercise, which, I try to do it consistently. And you know sometimes I don’t feel like doing it but I think exercising once a week, twice a week, three times a week — cardio specifically — to really get your heart rate up, I think that’s very important to keep a sharp mind. And then, also, I play a lot of video games too. You know now that the semester is over, I really enjoy video games. It’s a way to stay social and it’s a lot of fun so I say the two biggest things I do is game and workout.
Victor: What kind of games do you like?
Daniel: I kind of either like really competitive, group games like ‘Call of Duty: Warzone’, or this ‘League of Legends’ game called ‘Team Fight Tactics’. So, I really enjoy those. Then, I also enjoy single player like RPG kind of games. So, stuff like, ‘The Last of Us’, ‘Cyberpunk,’ those sorts of games. So it’s either like a competitive kind of group game or just a solo game where you get taken to another world. I played the new uncharted. It’s not new, but ‘Uncharted 4: A Thief’s End’ recently, and that game was a lot of fun.
AI and Gaming
“So what are your thoughts of AI and data science in terms of gaming?”
Daniel: Oh yeah, that’s a great question. Honestly, I think it’s interesting to see how it will manifest itself in games but I think the two biggest things are having non-player characters continually get more realistic with, not only neural networks controlling their behavior, but also things like giving them generated voices and generated faces to make them more realistic. And then also, having some sort of AI levels, AI scenarios where the game learns what the player is doing and then, on the fly, can generate new worlds or new levels. Stuff like that.
Christabelle: Yeah, that’s very interesting. I always think of how we differentiate between reality and dreaming and with all of the innovation, I think the two are beginning to blend together. It’s sort of this “what is real” and “what is dreamt.” I remember seeing this scary, virtual reality game and I think about how much trauma someone must go through if they experience that for the first time without any sort of pre-warning.
Daniel: Yeah, it’s crazy. I wonder how people who have never been exposed to like the internet would be to reacting some of these like VR horror games. It has a real psychological and physiological effect. And you’re absolutely right, I think a lot of things in our lives are designed to take us away from reality whether that’s like books — fiction books — or sports for some people. You know, you watch a game you kind of leave reality and just focus on this contest. So, it’s only going to get better and better. Pretty soon, you have VR games, AI games that are really going to test the boundaries of what we view as real versus not real.