Andrea G贸mez Jaime,听a first-year MSc student in the Biological and Biomedical Engineering听(BBME) program, recently听participated听in the inaugural听听(Hard Mode)听held at the MIT Media Lab from March 6鈥�8. In this student spotlight, Andrea shares what inspired her to apply, the project her team developed during the competition, and the lessons she learned from the experience.听
Q: Congratulations on your acceptance and participation in this hackathon! It must have been a really exciting opportunity. Can you tell us听abit听about听what the听MIT Hardware x AI Hackathon is[1] and what drew you to apply?
The Media Lab is well-established at MIT 鈥撎齣t's听an interdisciplinary lab with biomedical听engineering, design, computer engineering, and other听focuses 鈥� and this is the first time that听they鈥檝e听done this听hackathon听competition.听At 91丝瓜视频,听I鈥檓听part of the听听and the听, and this hackathon was related to both of these labs and the approaches we听use. At Hard Mode, the organizers wanted to reframe what a normal hackathon is, and what we can use software and applications for. With how AI听[artificial intelligence]听has re-shifted the types of things that we can do with the tools that we have, and the amount of things that we can accomplish in a short time, the organizers wanted us to see how we could embed听AI into the physical world.听The purpose was to build a physical product and not just an interface 鈥� which is what drew me to apply. I听thought this was a good challenge for me to reflect on how听I听could interact with AI outside of an app.听Now that AI is out there,听it鈥檚听not going to stop.听We will need听to听learn how to use it and what we can do with it听beyond applications and chat听bots.听
Q: What was the application process like?听
I heard about it by chance.听I鈥檓听a big fan听of the work done at the MIT Media Lab, and while I was visiting a friend in Boston, I came across the opportunity. It was perfect听timing. In the听application听process, I听had an interview and had to听submit听my portfolio of all the projects that听I鈥檝e听done across computer engineering, hardware, and wet lab-based听projects. I also had to听submit听a short paragraph about which track I wanted to听submit the project into, who I am, and what my interests are 鈥� this allowed for a pre-reflection of what I wanted to do in this competition and allowed me to reflect on how we could integrate AI into a new hardware and use it differently than how it鈥檚 already used.
Q: Can you briefly describe your research and how it connects to this competition?听What project did you work on during the听hackathon听and which of the 6 Core Tracks[2] were you involved in?
I鈥檓听working in computer vision systems that can read lateral flow assays and听then听I鈥檓听also able to work in public health through AI. In听the Shared Reality Lab, we听are encouraged to听have听side听projects that听we are part of external to our thesis, while we are听in the process of听completing听our听degree.听I鈥檝e听been听very interested听in the trend of smart patches and how听we can record biomedical data and听biosignals听from our bodies, other than with smart watches. In the lab,听I鈥檓听hoping to听work on a side project where听I鈥檓听interested in creating a smart tattoo!听听
Once I started building the team听within the hackathon, I knew I wanted to do something outside of my听grad听school research, and still use data tracking and a dashboard approach, integrating听vitals听measurements.听听
Our team chose the听搁别蹿濒别肠迟听迟谤补肠办 because many of our members had biomedical engineering听backgrounds听and wanted to use the data we collected in this competition for something other than health-related听projects.听We wanted something that would ground you in the moment, pulling you away from your screens听and allowing you to be more present. Ultimately, we designed a music box that generates music based on how your day was.听听
We mapped this听using听heart rate sensors and听galvanic skin response听sensors, tracking听changes听in skin听conductivity听associated with听micro-sweat听droplets听and听sweat gland听activity. The听micro-sweat听drops will be present when听you鈥檙e听excited or听stressed, but when听you鈥檙e听calm听or sad, their presence will be lower.听Similarly, heart rate听varies听as our听emotions听change, becoming faster or slower if we are more excited or calmer. We used听these signals听to听determine听and track听the user鈥檚听emotional state听throughout the day.听Then,听we used听an LSTM [Long Short-Term Memory network],听which is designed听to learn patterns across time,听to allow the system to understand how the user鈥檚 mood was evolving during the day and translate those patterns into musical parameters, like melody and tempo.听Based听on听this,听the system generates music that plays when the user hand-cranks the arm of the music box. The idea is that instead of just pressing a play button, users actively create their music as a grounding exercise.听
The music box is like听a听鈥淪potify Wrapped鈥� using your vitals!听
We听really took our time听conceptualizing听this听and it was great to work with people from other industries to see how they think and conceptualize ideas prior to creating the prototype.听
Q: How did your project turn out, and what challenges did you face during the hackathon?
We didn鈥檛 win, but it was still really fun听and I really enjoyed it!听I听hadn鈥檛听done a hackathon in a while, so I wanted to focus on having fun and experimenting with hardware, since I usually just code.听听
It was tough though 鈥� I worked on the sensor听portion听of the听project,听however, I听haven鈥檛听played around with sensors in 3-4 years.听I was working on connecting the heart rate sensor and galvanic听skin response sensors and sending the data through the AI pipeline, which one of my group members also worked on creating.听We听didn鈥檛听connect to an API听[Application Programming Interface]听from a wearable, but instead we wanted to create the sensors ourselves to allow us to collect raw data and听analyze it听how we wanted to,听in order to听generate the music. Wearables can have too many limitations because they provide processed scores, rather than raw data.听听
We also built 2 hardware devices, the听biosensors听and the music box. It was difficult because it was听a听lot听of work and working with biosensors is always challenging.听There is a lot of noise, and you have to be very precise with them.听It took longer than expected to extract the data听that I wanted because I had to ensure that the sensors were听accurate.听听
We used soldering in our听design,听but this was听fragile,听and听the soldering station was听located听separately听from听our team station. Because of this,听we had to ensure that we听didn鈥檛听break it, adding another element of difficulty.听听
We then had to understand how to generate the music 鈥� we understood what the vitals were telling us,听but how could we translate this into music? Since none of us had a background in music, we had to figure out these different metrics听and translate certain moods to various tempos and sounds. We听thankfully听have听some friends who听are musicians and music engineers,听so we called them and听we had them give us a听quick听masterclass in how this could work and how it could create music.听听
Finally, we wanted to incorporate the manual process of cranking the music box鈥檚 arm as a means of hearing the music. There听are听no mechanics built in that do anything when someone cranks听the arm, but instead sensors that听initiate听the music to play, simulating this action.听
Q: What was it like working and collaborating in such an intensive, fast-paced environment?
It鈥檚听stressful for sure.听I learned from past experiences that because it moves so fast you have to enter it with a schedule.听We spent听much听time brainstorming and building our听concept, but we听couldn鈥檛听compromise on this. It听was important that we created the concept and听stuck to听it. We had doubts that we had to talk through, but our main concept remained throughout and provided clarity for what we had to build.听
We also had daylight savings time taking place on the weekend of the hackathon, which听left people confused about timing and just added another hurdle.听
We had adrenaline building this though;听it was a lot of fun. The day before the deadline was the most stressful since the pitch is only 2 minutes long and we had to decide what we would say and the timing of this. Since I did 2 hackathons previously to this, I had experience, and I knew what I was getting myself into, which was important!听
Q: What new skills or perspectives did you gain from participating?听
I had never embedded AI into anything before 鈥� I have used models, but never put them anywhere, so that was cool to learn. I learned about hardware prototyping and using wearables with sensors as well... this was mostly a refresher of what I already knew,听but听I want to explore this in my master鈥檚 thesis听too听and it was good to review it. We also had people on our team from听different levels听of their听studies and听careers and different subject matter听expertise听鈥� so that was听very helpful听in learning about听how to conceptualize these kinds of designs and get varying perspectives.听听
Q: How do you see experiences like this shaping your future research or career direction?听
I think听it鈥檚听nice to spend time doing this. It seems school-related, but really, you have so much freedom to do what you want to do.听You鈥檙e听able to explore what you听usually听can鈥檛听inside the classroom.听Biomedical engineering is focused on health outcomes, which drew me to it, but听its听nice to听leverage听opportunities to explore other ideas that you听wouldn鈥檛听typically explore.听
I was outside of my usual environment, and I got to learn about other areas that I听didn鈥檛听know as much about听or things that I put aside for a little while.听It鈥檚听nice to see what other things you can do with the skills you have.听When you graduate, you have to be flexible.听You may not end up in the biomedical industry听as an engineer, so听it鈥檚听important to be able to apply your skills elsewhere.听听
Q: What advice would you give to other biomedical engineering and听bioengineering trainees听interested in听applying听to听similar opportunities?
You should definitely do it!听It鈥檚听a lot of work, but I听don鈥檛听regret spending my reading week doing this.听It鈥檚 a good听opportunity听to connect back to why you joined engineering in the beginning.听Coming听into听the competition with a general idea听of what you want to build听or听a听concept saves you time and really makes a听difference in the final听product.听Understanding听peoples鈥櫶齨eeds is also an important part of biomedical engineering;听you鈥檙e听doing things because听they鈥檙e听fun, but also because they have a purpose!听
听
Thank you, Andrea, and best of luck with your future projects!听
听
听
听
[1]From the website:听鈥淗ARD MODE is MIT's 48-hour hardware 脳 AI hackathon focused on building the future of intelligent objects: devices that sense, learn, adapt, and respond to the people around them. The challenge is to imagine what else AI could be. Not another chatbot. Not another app. Real hardware you can hold, wear, share, install, and live with.听200 participants will integrate AI with physical systems to prototype tangible, functional artifacts. Systems that rethink how humans connect, learn, reflect, work, play, and thrive. Hosted by MIT Media Lab's AHA and the Design Intelligence Lab, HARD MODE brings together engineers, designers, and researchers to push the boundaries of what intelligent hardware can be.鈥澨�
[2]From the website:听鈥淧LAY:听Create joyful hardware: interactive stories, hybrid games, AI-driven performance tools, and installations that spark wonder and shared play.听LEARN:听Reimagine how people learn with adaptive devices, generative study aids, and tools that turn abstract ideas into accessible, hands-on experiences.听WORK:听Invent AI-powered tools and processes that expand human capability: new mediums, generative hardware, and systems that help people create the impossible.听CONNECT:听Create tech that strengthens relationships and听community-devices听that reduce loneliness, support organizing, and help people find belonging.听REFLECT:听Design technologies that support contemplation and听growth-tools听for mindfulness, mood tracking, habit reflection, and understanding yourself.听THRIVE:听Build tech that boosts performance, care, and autonomy: from wearables and assistive tools to systems that help people live longer and better.听鈥澨�
听