Creative Lens AI
Creative Photo Assistant
Engineered a real-time multimodal AI assistant to eliminate the creative friction of photography and content creation. By integrating Gemini 2.5 Flash with a Python/Streamlit architecture, the app performs instant visual analysis to provide context-aware posing tips, AR lens recommendations, and trendy captions. This tool streamlines the journey from 'camera-shy' to 'post-ready,' allowing users to focus on the moment rather than the stress of getting the perfect shot.
Personal Project/Open Source
Artificial Intelligence/Augmented Reality (AR)/Creative Technology/Social Media & Content/Consumer Software
Full-Stack Development/AI Engineering & Prompt Design/Product Design (UI/UX)/Computer Vision Integration
February 2026
The Build
To ensure a high-performance, mobile-first experience, I architected a Multimodal AI pipeline using the following stack:
Logic & Integration: Developed in Python 3.12, utilizing the new Google GenAI SDK for seamless communication with the Gemini API.
AI Model: Integrated Gemini 2.5 Flash-Lite to balance high-speed image reasoning with strict rate-limit management, ensuring the app remains responsive under load.
Interface: Built with Streamlit, customized via CSS injection to create a branded, "Snap-style" UI that supports both live camera streams and local file uploads.
Security: Implemented Environment Variable (Dotenv) protection for API credentials and established a robust Error Handling system with exponential backoff to manage API 429 (Resource Exhausted) quotas.
Deployment: Managed through GitHub with a continuous deployment pipeline to Streamlit Community Cloud, utilizing encrypted secrets for secure production access.
Technical Specifications:
Multimodal Logic: Integrated Gemini 2.5 Flash-Lite to process both text and image data simultaneously for real-time visual reasoning.
Robust Error Handling: Engineered a retry mechanism with exponential backoff to gracefully manage API rate limits (429 errors) and ensure 99% app uptime.
Custom UI/UX: Utilized Python and Streamlit with injected CSS to build a responsive, branded interface featuring custom camera integration and file upload capabilities.
Security-First Architecture: Implemented industry-standard security protocols using .env files and Streamlit Secrets to protect API credentials.
