Atom

• 

  Atom AI Logo

Inspiration

Caleb and I saw firsthand how time-consuming and frustrating employee scheduling could be. Managers often spent hours crafting weekly schedules—only to face shift conflicts, employee dissatisfaction, and last-minute changes. Despite their best efforts, the results rarely met both business demands and employee preferences. We knew there had to be a better way. That’s when we imagined an AI-powered tool that could simplify the entire process—creating smarter, faster, and more balanced schedules for everyone involved.

What it does

Our AI-powered scheduling app intelligently creates optimal work schedules by balancing business requirements with employee availability and preferences. It automates the entire scheduling process-factoring in variables like peak business hours, shift coverage, labor laws, time-off requests, and employee roles. Managers can generate, adjust, and publish schedules in minutes instead of hours. The result is a more efficient, fair, and flexible schedule that meets both operational demands and employee satisfaction.

How we built it

We used a combination of AI tools and our prior experience in programming, database design, and real-world problem solving to plan and develop the foundation of the app. Drawing from our firsthand experience in the service industry, we identified the key pain points in manual scheduling and designed a system architecture to address them. We used backend technologies to handle data storage and logic for schedule generation, while AI algorithms help evaluate optimal shift combinations based on multiple constraints. Our troubleshooting skills were essential throughout the process as we iterated through different approaches to create a smart, efficient, and user-friendly scheduling solution.

Challenges we ran into

One of the biggest challenges we faced was training a machine learning model to improve and adapt its scheduling decisions over time. While we were familiar with programming and data structures, applying machine learning in a real-world context—especially for something as dynamic and constraint-heavy as employee scheduling—was new territory. We had to navigate issues like handling incomplete data, balancing conflicting priorities (business needs vs. employee preferences), and tuning the model to recognize patterns that result in more effective schedules. Learning how to train and evaluate the model's output was a steep but rewarding learning curve.

Accomplishments that we're proud of

The machine learning model that learns from previous schedules metrics and adapts is a very cool and useful feature that I think businesses will benefit from.

What we learned

We learned about machine learning models, setting metrics that can be used to adjust the weights and biases and seeing the ways that the model benefits or worsens from that development.

What's next for Atom

We are going to create a login system that allows our application to be used by multiple companies. We also would like to train the model on many more schedules and work towards creating the most optimally trained model for schedule creation.

Built With

• bolt
• npm
• sqlite