Implementation proof
The paid cases come first. They show the failure, my role, the shipped change, and the verification method. The engineering projects show the systems behind that work.
Paid client work
Client identities stay private. Scope, implementation ownership, and validation remain specific; business outcomes are not invented.
Paid client delivery
A paid route and release QA engagement for a headless Next.js/WordPress build, separating what could be diagnosed in generated output from what had to be verified on the immutable and production deployments.
Paid client delivery
A paid recovery engagement for a WordPress site where valuable legacy paths needed explicit owners, safe redirects, and live verification rather than a blanket redirect rule.
Anonymized paid dental-sector work
Historical multi-milestone work for a dental practice, with redirect, crawl, entity/schema, rendered business-data, Google Ads conversion-tracking, and technical QA outputs kept separate from any ranking or conversion outcome.
Anonymized paid professional-services work
A completed multi-month professional-services engagement where live GA4/GTM funnel measurement and GSC/GA4 reporting were implemented and handed off without treating tracking deployment as proof of traffic, subscriber, lead, or revenue growth.
Reseller-delivered paid work
A bounded WordPress implementation delivered through a reseller relationship, with Niko responsible for the agreed technical changes and live checks while the reseller retained the end-client relationship.
Built systems
Code, interfaces, and explicit maturity boundaries—not decorative portfolio cards.
Open-source technical SEO CI gate
An open-source Rust CLI that turns static output or a live site into one URL-state graph, reports deterministic findings when crawl and indexability signals contradict each other, and fails CI according to the configured severity gate.
Open-source PostgreSQL operations tool
An open-source Go reconciler for pgvector HNSW indexes that belong to a defined workload or filtered population. It detects drift, plans the required change, builds and verifies a replacement concurrently, publishes it under the desired name, and retires only indexes from the same managed family.
Open-source PostgreSQL diagnostic
An open-source Go CLI for testing a failure that latency-only vector benchmarks can miss: a filtered approximate search can lose relevant neighbors, and a seemingly good result can hide a sequential-scan fallback. The tool compares ANN results with exact PostgreSQL ground truth and checks the measured query plans.
Private crawler / audit evidence
ILCrawler turns an authorized site audit into durable page, link, resource, render, performance, content-similarity, and issue evidence, then keeps review and exports attached to the run.
SEO decision layer
Internal system for deciding whether an SEO task is worth doing before it becomes a ticket.
Search intelligence cockpit
Private dashboard that connects a site, its crawl evidence, market research, and approved SEO actions in one place.
Workflow / outreach ops
Single-owner outreach system for tracking targets, messages, replies, bounces, forms, follow-ups, and dedupe protection.
Evidence limits
A correct redirect, canonical, route response, or tracking event can be verified immediately. Rankings, traffic, and revenue require later observation and are not invented here.
