Freelance AI Engineer, Media Monitoring System

About the role

Gold PR, a PR and consulting firm in Tel Aviv, is building PR Radar, an in-house AI system that will track every mention of our clients, prospects and fields of interest across X, YouTube, global news wires, Telegram channels and court filings, in Hebrew, Russian and English, 24/7. Every item is read by an LLM, scored for relevance, newness, source weight and momentum, clustered into stories and delivered as a live board, a daily 08:00 digest and push alerts when something urgent breaks.

The spec and the design are done. We need the engineer who turns them into a working system and stays on to run it.

What you’ll do

• Own the full pipeline, from collection (X, YouTube, news, Telegram, court sources) through de-duplication, tagging and storage
• Build and tune the LLM layer that scores every mention for relevance, stance, source credibility and newness, with every decision explained
• Develop the clustering that turns dozens of posts into one story, finds the original source and tracks developments without repeating old news
• Build the alerting and delivery side: three-tier importance model, live board, daily digest, push notifications
• Keep it reliable and cost-controlled, and work directly with the project owner and PR team to turn feedback into features

What we’re looking for

• Working knowledge of Hebrew, Russian and English. You’ll be reading the sources the system reads, and judging whether it got them right
• 4+ years building production data pipelines or backend systems
• Hands-on experience with LLM APIs (OpenAI, Anthropic or similar) in production, including prompt design and evaluation at scale
• Solid understanding of scraping, rate limits, queues and scheduling
• Ability to build a clean, functional web interface
• Has built a system from zero alone and run it in production