AI-powered fitness apps are changing personalized training by combining data, recovery insights, and adaptive workout planning into a single experience designed around individual goals.
For Manikanta Sirumalla, UMBC Data Science graduate student, that realization came from years of personal experience in the gym. After struggling with disconnected fitness apps and generic workout plans, he decided to build something better.
In this conversation with Dennise Cardona, M.A. ’23, Manikanta reflects on his journey from iOS developer to founder, how UMBC shaped his thinking, and what it takes to build an AI-powered fitness platform from the ground up.
Building Smarter AI-Powered Fitness Technology
Dennise Cardona:
What led you to pursue graduate studies in data science at UMBC?
Manikanta Sirumalla:
I’m originally from India, and before coming to UMBC, I worked for about three years as a professional iOS developer.
During that time, I worked on many different mobile applications where data played a huge role. That’s when I started becoming more interested in AI-powered systems and data science.
I realized the future of mobile applications would be heavily driven by AI-powered personalization and intelligent systems. I wanted to understand that world more deeply and build AI-powered applications that could actually learn and adapt to users.
When I started researching universities, UMBC stood out because of its strong focus on technology, innovation, and opportunities like BWTech. That’s what brought me here.
The Problem With Existing AI-Powered Fitness Apps
Dennise Cardona:
You’ve talked about your personal fitness journey inspiring Rep Track Pro. What was missing in the tools already out there?
Manikanta Sirumalla:
I’ve been training for around 10 years now, and when I first started, I had no idea what I was doing.
I was trying to figure everything out while using different AI-powered fitness apps, but the biggest issue was that none of them worked together.
Your calorie tracking stays in one app. Your workouts stay in another. Recovery data is somewhere else entirely.
I found myself manually copying information between apps just to plan my week properly. At some point I realized the tools were supposed to help me, but instead I was working for them instead of using a true AI-powered system.
That frustration became the starting point for Rep Track Pro, an AI-powered fitness platform.
The Moment the Idea Became Real
Dennise Cardona:
When did you realize this could become more than just a side project?
Manikanta Sirumalla:
It happened gradually.
One moment that really stayed with me was seeing someone at the UMBC gym tracking workouts with a pen and paper.
I asked him why he wasn’t using an AI-powered fitness app, and he told me, “All the apps I’ve tried are garbage.”
That really hit me because I felt the same way about most AI-powered fitness tools.
I realized this wasn’t just my personal problem. A lot of serious fitness users felt disconnected from AI-powered tools available to them.
That’s when I started thinking bigger: What if there was one AI-powered system that connected workouts, nutrition, recovery, AI coaching, and analytics together in a meaningful way?
As a data science student and iOS developer, I felt like I finally had the right background to build an AI-powered solution.
Building an AI-Powered App as a Solo Founder
Dennise Cardona:
What were the biggest challenges of taking the app from an idea to a live product?
Manikanta Sirumalla:
Time management was probably the biggest challenge.
I started building the AI-powered app during my first semester, but by my second semester I was balancing three core graduate courses, multiple projects, and AI-powered app development at the same time.
There were months where I was sleeping only three or four hours a night trying to keep everything moving.
Then there was the App Store approval process for the AI-powered app. My app was rejected multiple times before it was finally accepted.
And even after launch, the work didn’t stop. Users would report crashes or issues in the AI-powered system late at night, and I’d immediately have to investigate and fix them.
But honestly, all of that helped me grow—not just technically, but mentally as a founder building AI-powered technology.
Building Practical AI-Powered Fitness Intelligence
Dennise Cardona:
Your app uses AI for workout and nutrition planning. How did you approach building that in a practical way?
Manikanta Sirumalla:
One thing I noticed about many AI-powered fitness apps is they simply send your information to an LLM and return whatever response comes back.
That creates problems because AI can invent exercises, ignore available equipment, or give nutrition advice that isn’t grounded in real science.
I didn’t want Rep Track Pro to be just another AI-powered fitness app like that.
So I built what I call a deterministic AI-powered architecture.
The calculations for calories, protein, recovery, and body metrics are all based on evidence-backed formulas. Then AI helps personalize recommendations using the AI-powered exercise database I created.
For example, the app has over 2,000 exercises stored in the system. AI selects the best exercises based on your goals, available equipment, and training style within an AI-powered framework.
The app also includes instructional videos for every movement so users don’t need to leave the AI-powered ecosystem to search online.
I wanted everything to exist in one AI-powered fitness ecosystem.
Bringing Classroom Learning Into AI-Powered Real-World Systems
Dennise Cardona:
How did your coursework at UMBC shape the app?
Manikanta Sirumalla:
The biggest impact came from my machine learning fundamentals course.
Before that class, my understanding of AI was mostly limited to chatbots and AI-powered language models.
But learning about regression models, classifiers, and machine learning systems completely changed how I thought about building AI-powered applications.
I actually went directly to my professor and explained what I was building. He helped guide me toward the right AI-powered models and approaches. UMBC has a great support system and helps students achieve great success.
Right now, Rep Track Pro uses AI-powered regression models for weight prediction and AI-powered classifiers for recovery analysis.
I’m also taking project management courses that are helping me think beyond development and focus more on scaling an AI-powered product.
Winning the Cangialosi Business Innovation Competition
Dennise Cardona:
You recently won first place at the Cangialosi Business Innovation Competition. What did that experience teach you?
Manikanta Sirumalla:
The biggest lesson was learning how to tell an AI-powered story clearly and defensibly.
Early on, I focused too much on features instead of focusing on what I could confidently explain and defend in an AI-powered context.
I realized judges aren’t impressed by a long feature list. They care about whether you truly understand your AI-powered system and can explain its value under pressure.
I also learned the importance of framing an AI-powered product.
At first, I viewed being a solo founder and pre-revenue startup as weaknesses. But later I realized those could actually become strengths if presented correctly in an AI-powered narrative.
Expanding AI-Powered Personalization in Fitness
Dennise Cardona:
One feature that stands out is the app’s menstrual cycle-based training support. What inspired that?
Manikanta Sirumalla:
Initially, I wasn’t sure whether I should include it because I didn’t want to make assumptions about AI-powered recommendations.
So instead of guessing, I created surveys and gathered feedback from women across different universities and fitness communities.
That research helped shape the AI-powered feature.
The app adapts workout intensity, recovery expectations, and nutrition suggestions based on different phases of the menstrual cycle within an AI-powered system.
For example, during phases where recovery may be lower, the AI-powered platform adjusts training recommendations and suggests nutrients like iron-rich foods.
I wanted the feature to be grounded in real experiences and real feedback, not just AI-powered assumptions.
Looking Ahead
Dennise Cardona:
What’s next for Rep Track Pro and for your future after UMBC?
Manikanta Sirumalla:
The next big milestone is launching the premium AI-powered version of the app.
I’ve also applied to the UMBC Launchpad accelerator program because I want to continue growing Rep Track Pro beyond campus and eventually expand this AI-powered fitness platform to universities and fitness communities everywhere.
I’m even exploring opportunities with Y Combinator for this AI-powered startup.
This experience has taught me so much—not just technically, but personally and professionally through building AI-powered systems.
I genuinely feel like building this AI-powered product while studying at UMBC accelerated my growth in ways I never expected.
Building the Future of AI-Powered Fitness Apps
For Manikanta, Rep Track Pro represents more than just an AI-powered fitness app.
It’s the combination of personal experience, technical expertise, data science, and entrepreneurship—all coming together in an AI-powered system to solve a real-world problem.
And as AI-powered technology continues shaping the future of health and wellness, his work reflects a larger shift happening across industries: the move toward systems that are smarter, more connected, and truly AI-powered for real human needs.
UMBC Data Science Graduate Program
Learn more about how UMBC’s Data Science program can help you achieve career success.