dnurvianto
I am a Full-Stack Systems & Technical Search Engineer (15+ years experience) with a formal background in Statistics (B.Sc., Universitas Gadjah Mada). I help high-growth SaaS and enterprise platforms bridge bare-metal infrastructure with generative search visibility (AEO/GEO).
Most AEO/GEO efforts fail because they treat LLM discovery as a surface copywriting problem. In reality, generative engines (ChatGPT Search, Perplexity Sonar, Google AI Overviews, Claude) require a continuous 3-tier computational continuum:
1. Tier 1: Upstream Server Infrastructure — Tuning origin Linux/Nginx stacks and edge caches to deliver sub-80ms TTFB. This eliminates Host Load Throttling so AI scrapers (GPTBot, PerplexityBot, ClaudeBot) crawl deep documentation without timeout queues. Tuning database buffer pools (60-70% RAM) and Redis object caches over UNIX sockets to prevent dynamic SSR compute leaks.
2. Tier 2: Web Architecture & Data Ingestion — Eliminating heavy DOM payload bloat (<800 nodes, zero-CLS containment) and sandboxing third-party script hydration via Web Workers. Deploying partitioned XML sitemap indexes, clean /llms.txt streams, and dense, multi-entity JSON-LD graphs (SoftwareApplication, Organization, E-E-A-T triples) formatted directly for LLM RAG pipelines.
3. Tier 3: Downstream On-Page AEO & Telemetry — Structuring semantic token dispersion using deterministic 5W+1H answer blocks and strict topical silos ("1 Pillar, 10 Clusters") to prevent AI summarization drift. Using mathematical variance modeling to objectively quantify prompt share-of-voice and citation shifts.
Available for technical audits, infrastructure sprints, or white-label technical execution partnerships. 100% remote async, zero meeting overhead, W-8BEN compliant (zero US tax/payroll friction).
Feel free to reach out directly at [email protected].