🏆 Third Place Overall, MLH SharkByte Hackathon 2025

lunarfeed

Digital Radio News Broadcast

AI-driven digital radio news broadcast featuring an automated backend pipeline that scrapes, aggregates, and synthesizes space and science news into character-driven audio broadcasts.

Role Backend Engineer
Achievement Third Place Overall
Event MLH SharkByte Hackathon 2025
Visit Live Site →

Overview

lunarfeed is an innovative digital radio station that brings space and science news to life through AI technology. As the backend engineer, I engineered and automated the complete pipeline that powers this unique broadcast experience, transforming raw news data into polished, character-driven audio content.

The system operates autonomously, continuously monitoring multiple RSS feeds, generating broadcast scripts with fact-checking guardrails, and producing realistic audio that matches a 1950s-style radio host persona.

Tech Stack

Python AWS Gemini API ElevenLabs API RSS Automation Scripting

Key Features

  • RSS Feed Aggregation

    Developed Python scripts to scrape and aggregate news from NASA and other space/science RSS feeds, ensuring comprehensive and up-to-date coverage of the latest developments in space exploration and scientific discovery.

  • AI-Powered Script Generation

    Integrated Gemini AI API to automatically generate broadcast-ready news scripts. Implemented fact-checking guardrails to guarantee factual accuracy against original sources, maintaining journalistic integrity throughout the automated process.

  • Character-Driven Audio Synthesis

    Leveraged ElevenLabs API to synthesize realistic, character-driven MP3 audio readings of news scripts. Fine-tuned the voice synthesis to match a 1950s-style radio anchor persona, creating an immersive and nostalgic listening experience.

  • Cloud Infrastructure & Streaming

    Deployed AWS infrastructure for archiving audio files and enabling fast, reliable front-end streaming for listeners. Optimized storage and delivery pipelines to ensure seamless playback across devices and network conditions.

System Architecture

The backend pipeline operates as a fully automated system:

  1. Data Collection: Python scripts continuously monitor RSS feeds from NASA and science news sources, extracting relevant articles and metadata.
  2. Content Processing: Gemini AI analyzes and synthesizes content into broadcast-style scripts while maintaining factual accuracy through validation against source material.
  3. Audio Generation: ElevenLabs API converts scripts into MP3 audio with consistent voice characteristics matching the radio host persona.
  4. Distribution: AWS handles storage, archiving, and streaming infrastructure, ensuring reliable content delivery to end users.
sequence diagram depicting the backend automation sequence
Backend automation sequence diagram

Impact & Recognition

lunarfeed earned Third Place Overall at the MLH SharkByte Hackathon 2025, recognized for its innovative integration of AI technologies, robust automation pipeline, and creative approach to news broadcasting.

The project demonstrates the potential of AI to transform traditional media formats while maintaining editorial standards and factual accuracy. It showcases my ability to architect complete backend systems, integrate multiple APIs, and deliver production-ready solutions under tight deadlines.

← Back to Projects