PocketTrainer
PocketTrainer is an AI-powered, cloud-hosted fitness app that puts a personalized fitness trainer, in your pocket.
PocketTrainer is a fitness app that analyzes a gym user's workout form by comparing their workout footage to our trained AI model. Our project is using bubble.io to quickly create and deploy a web-based app and is leveraging the Peltarion AI model to analyze and correct a user's form in real-time. The app will rely on users recording their workouts with their smartphones in the appropriate angle and lighting and will be able to receive feedback on how to improve their workouts.
We'll be lowering the barrier of entry of expensive personal trainers. About 87 million American adults go to the gym, spending ~$34B on memberships annually. Most gyms lose 49% of their new members within 6 months. Gym users often blame ‘being confused as to what workouts to do, having difficulty being consistent, feeling aches after performing improper exercise form, and making unhealthy eating choices.’ We want to change that.
Corrects your form in real-time
Your form is compared to a human model based on your body type, structure, and height.
Instructs users on how to perform workouts correctly with real-time feedback.
Ex. Adjusting hips, knees, and feet while squatting
Tracks your workouts, sets, reps, and weight amounts to help users log workout data → data-driven workout recommendation improvements.
In the app's future, we plan on incorporating these features:
Your workouts will be tailored to your fitness goals
Building muscle mass
Losing body fat
Create sweepstakes bet with friends to keep each other accountable at the gym every week
Up to 6 people per party
Party participants compete weekly
Users that hit their goals have the prize amount divided amongst the winners
See more of our development process here
We have created a Google Form to perform customer interviews and create a beta email list for when we release our product. We have performed 20+ interviews as of November 2021.
Additionally, we have spoken to various gym users and 3x gym influencers/powerlifters (100k+ followers in total) about how we should develop the algorithm and what to focus on. We also have the support of two personal trainers from our local gym.
- Improving healthcare access and health outcomes; and reducing and ultimately eliminating health disparities (Health)
- Concept: An idea being explored for its feasibility to build a product, service, or business model based on that idea
We have built out a prototype during the ShellHacks hackathon. We have pivoted from using Flask and OpenCV to a no-code solution in order to go to market quickly. We are using bubble.io and plan on incorporating it with Peltarion AI to train our fitness models. Then we will get our domain name and make the project public.
- A new use of an existing technology (e.g. application to a new problem or in a new location)
The core technologies behind our application are cloud computing for real-time feedback, and artificial intelligence to compare user workout form to a model.
Our core technology needs a medium to be distributed, in our case, we are using bubble.io to create a no-code application and users would be using their smartphone cameras to record workout footage.
- Artificial Intelligence / Machine Learning
- Software and Mobile Applications
- United States
Our application is currently in its pre-beta stage, so we have zero users being served. However, we've begun potential customer interviews with people, mostly ages 18 - 35 at our local gyms and on campus.
As our application's beta becomes available and we are ready to scale, we will serve 1,000+ users in the coming year.
This application will be used directly by new gym users, amateur gym users, and or personal trainers with their clients to scale their reach. This solution will lower the barrier of entry for gym users to progress in their weight lifting journey and drastically reduce the risk of injury from faulty workout forms.
Our impact goals are simple: improve workout form, enable faster weight load progression, and reduce risk of injury during workouts.
We plan on achieving these goals by training our AI model with footage from top fitness athletes and bodybuilders, test and iterating user input for ideal camera angles, and getting feedback from personal trainers and powerlifters to develop a product by fitness enthusiasts, for fitness enthusiasts.
On top of the direct functionality aspect of our app, we will also be influencing our community by lowering the financial barrier of entry for gym users. Personal trainers can cost dozens, if not hundreds of dollars per week. Our app would be free to use at first, and function on a subscription model, with an expected price of $5 per month.
We will break down our measurements into different buckets: app development progress, AI model accuracy, number of users using the app per month, and number of customer interviews/engagements our team has per month.
App development progress will be measured by the number of app pages created, the cleanliness of the UI/UX, the functionality of all the buttons and data storage.
AI model accuracy will be measured by getting feedback from powerlifters and personal trainers that will be able to help full-proof the AI model.
The number of users per month will be taking a product management approach, as we can analyze our churn rate, frequency of use per user, and the number of referrals per user.
Lastly is our customer interviews and engagements. This is everything from probing the problem we are solving, feedback on the beta app, and iterating based on the feedback we attain from the fitness market.
Our current barriers are:
Intellect - Having our team learn how to build a functioning app, teach an AI model, and give in-app workout feedback
Market - Testing to see if anyone wants the solution. We have positive feedback from potential users and colleagues, but we won't know until we bring a beta product to market.
Technology - Testing if the AI model and camera would come out with accurate form feedback. We are not 100% confident the technology will work well until we test it out.
Financial - Having enough runway to cover our software development costs while we iterate our solution, marketing efforts to scale and acquire users, paying for professional input from personal trainers and powerlifters to ensure it's a high-quality solution, and reimbursements for customer interviewers' time.
Our team has a passion for fitness, reaching our goals, and making a scalable solution. Our goal is to reduce the barrier of entry for more people to live healthier lives. Our team is well-positioned to create this app with our diverse skillsets.
Branden Bedoya - Branden's expertise lies predominantly within Business Development, Data Analytics, and Finance. He is a Business Analytics and Management Information Systems major at FIU, has finance experience working at Citigroup, Google, and soon IBM, and has participated in other pitching events such as Hult Prize. Within the team, Branden acts as our no-code developer, UI/UX designer, and project visionary.
Zack Stines - Zack's expertise lies within content creation, marketing, design, and operations. He has experience creating personal skits and corporate content on Tik Tok, Instagram, and Facebook. He has worked in digital marketing, and currently handles operations and logistics within his painting business.
Akash Samal - Akash's expertise is within research and software development. Akash led our team's prototype created during the Shellhacks hackathon, incorporating OpenCV on Flask. Akash has experience creating a racing-type game on the Google Play Store and has conducted undergraduate research at Lockheed Martin.
Rose De Sicilia - Rose's expertise is within science - notably nutrition and neuroscience, software engineering, and acting as a business development consultant. Rose learns from researchers like Andrew Huberman in her free time, is an undergraduate computer science student, and has worked alongside Branden within SoftBank Operator School's course, laying the framework for the app's PR FAQ.
We currently partner with staff from StartUP FIU. In particular, we have received feedback from:
- Gustavo Grande with the Rapid Innovation Studio on framing our problem and approaching our solution using design thinking.
- Robert Hacker - the co-founder of StartUP FIU has assisted in approaching how to perform customer interviews.
- Michael Burtov has assisted with giving insights on corporation structuring, scaling customer interviews, and general startup insights.
SoftBank Operator School - We have been able to gain feedback and resources from staff
- Laura Gaviria Halaby - as a general resource to tap into SoftBank's network.
- Daniela Cadena - on developing our PR FAQ and probing us to reframe our problem.
- No
NA
- Yes
Impacting women's lives: boosting confidence, improving lifestyles, lowering the barrier of entry to reach fitness goals.
We will use the prize to advance our solution by expanding our team, expediting our app development progress, performing more customer interviews at scale, and begin our marketing costs.
Student