System StatusOnline
NAV
GiftFinder is an advanced recommendation engine designed to solve consumer decision paralysis. Unlike standard catalogs, this is an autonomous agentic system that analyzes the market, discovers products, ranks them, and generates unique SEO content for over 11,000 intent permutations (e.g., 'Gift for a programmer for a housewarming under $50'). The goal: dominate long-tail e-commerce keywords with zero manual editorial work.
Optimized for cost and throughput. I use a cascade model where lightweight models handle logic and heavy models handle creativity.
Generates a logic matrix for URLs: /recipient/occasion/budget/interest. Includes a custom 'Logic Matrix' that prevents nonsense combinations (e.g., blocking 'Alcohol' for 'Child'). Implements full JSON-LD Schema.org graphs.
Instead of static XML imports, an AI Agent mimics a human researcher. It brainstorms 50+ niche queries per category, validates them against live API data, and applies a custom 'GiftScore' algorithm to reject low-quality items.
Uses Semantic Deduplication (Jaccard Index) to avoid repetitive products. The AI selects a 'Top 19' list based on the specific persona and writes a unique 'Why this works' justification for every single item.
A custom admin dashboard for queuing jobs, prioritizing seasonal trends (e.g., 'Christmas'), and a 'Blog Automator' that identifies viral topics and writes HTML-formatted articles.
The engineering capabilities applied in this case study are available as standalone services.
'LLM-as-a-Judge' architecture for automated brand visibility and sentiment scoring across ChatGPT, Gemini and Perplexity.
View case studyFintech / AISmart SPA Fallback architecture combined with Python Serverless and 90% editorial automation via Gemini AI.
View case studyBusiness IntelligenceInternal BI dashboard aggregating SEO, Ads and work logs into a real-time profitability engine with AI insights.
View case studyWhen users ask ChatGPT instead of Google, the rules change. Discover GEO — the engineering of visibility in the age of language models.
9 minAI & AutomationAI implementation starts not with choosing a tool, but with identifying one repetitive process that wastes the most human time. Learn step by step how to select, map, and automate that process.
8 minInitiate protocol. Establish connection. Let's build something loud.