Align

Inspiration

Making holistic wellness simple. Many people struggle to maintain a balanced lifestyle. The current market is filled with fragmented wellness tools — separate apps for journaling, workouts, meditation, and health tracking. This fragmentation makes it difficult to stay consistent, track progress, or gain meaningful insights about one’s well-being.

Our solution: Align integrates personalized recommendations, emotional reflection, and AI-powered physical and health analysis into one unified platform — empowering users to achieve balance in both body and mind.

How we built it

Our stack is fully serverless and designed for scalability and personalization:

Frontend - We used React Native to build clean, responsive interface that encourages reflection and consistency.

Backend:

AWS Lambda (Node.js and Python) functions handle journaling, sentiment analysis, and AI-driven emoji generation

DynamoDB (NoSQL) stores user journals, metadata, and personalized insights efficiently. We also used Amazon Bedrock to power our text intelligence and chat interface.

AWS Rekognition - Analyze workout form and give feedback to the user.

AWS Sagemaker - Video analysis with facial tracking and signal processing

What we learned

Unify data and AI workflows using AWS Lambda, API Gateway, and DynamoDB, allowing seamless data flow between journaling, workout analysis, and health tracking.

Leverage Amazon Textract for OCR to process handwritten journals and convert them into searchable text.

Use Amazon Comprehend for NLP and sentiment analysis, enabling emotional reflection and mood tracking from users’ written entries.

Apply Amazon Bedrock for AI text generation, powering journal summarization, emoji generation, and smart journaling suggestions.

Implement Amazon Rekognition to perform computer vision analysis on workout videos, detecting arm position and orientation with custom-trained labels.

Train and deploy SageMaker models to run remote photoplethysmography (rPPG) pipelines, extracting users’ heart rate from facial video by analyzing subtle skin tone variations.

Use EC2 instances for heavy computation tasks like frame splitting, data preprocessing, and testing large ML pipelines before deployment.

Accomplishments that we're proud of

Seamless AWS Integration: We successfully connected nine AWS services (EC2, SageMaker, DynamoDB, Lambda, API Gateway, Rekognition, Textract, Comprehend, and Bedrock) into one cohesive system.

Real-Time rPPG Heart Rate Detection: Implemented a camera-based, non-contact heart rate estimator using SageMaker and MediaPipe — capable of detecting subtle color changes on the face with FFT signal processing.

AI-Augmented Journaling: Used Bedrock and Comprehend to enhance journaling with sentiment detection, emoji tagging, and emotional summaries — transforming plain text into insightful self-reflection.

Workout Motion Tracking: Trained a Rekognition Custom Labels model on 200+ annotated frames to analyze arm orientation and provide instant feedback on workout form.

Unified User Experience: Integrated journaling, emotional insights, and health analysis into a single platform — achieving our mission of making holistic wellness simple.