SEO & Core Web Vitals Engineering
Engineering for how search engines — classic and AI — crawl, parse, and index the site. Schema, sitemaps, LCP/INP/CLS, AI-bot policy.
This is the technical SEO engagement. The deliverable is the technical SEO surface engineered as code in the buyer's repository, not a PDF audit that lives on a shelf. JSON-LD schema graphs that validate. Database-driven sitemaps that regenerate on publish. Canonical and redirect architecture defined in the application. Core Web Vitals budgets enforced in CI. Named AI-bot policy in robots.txt with one line per crawler.
I do not promise rankings. Search results are owned by the search engine, not by the engineer fixing the plumbing. What I promise is the artifacts — schema graph, sitemap generator, 301 map, Core Web Vitals dashboard — and the measured before-and-after on the signals the engine actually consumes. The parent — Website Creation & Redesign — sets the broader perimeter; this page is the focused SEO and Core Web Vitals workstream.
Schema graphs
JSON-LD generated from the application's actual data — Organization with sameAs links to LinkedIn and GitHub, Article with a real Author node, BreadcrumbList, FAQ, Service, LocalBusiness where applicable. Validated against the Rich Results Test on every deploy. Schema regressions alert. The schema is connected — a single graph that nests properly, not five disconnected blocks that confuse the parser.
Sitemaps
Multi-sitemap index. Generated from database queries, not hand-edited. Low-value pages filtered at the SQL layer rather than generated and then hidden. Priority and changefreq tuned per content tier. Lastmod updated on publish, not stamped on every regeneration.
Canonical and redirect architecture
A written canonical policy across www / non-www, http / https, trailing-slash variants, query-string variants. Hreflang for multi-locale sites that need it. The 301 redirect history preserved through every replatform and every URL change. The redirect map lives in code, with tests.
Core Web Vitals
LCP, INP, and CLS budgets enforced in CI. The build fails if a deploy regresses the budget on a representative page set. Field data pulled from the Chrome User Experience Report monthly to verify what real users see, not just what Lighthouse measures in a clean room. Diagnoses come with code fixes, not "consider optimizing your largest contentful paint."
AI-bot policy in robots.txt
Named per crawler. GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, Google-Extended, PerplexityBot, Applebot-Extended, Bytespider, CCBot — each with its own line and its own policy (allow, throttle, or deny). Wildcard rules are ignored by most AI bots; the policy has to be explicit per user-agent. The policy is chosen with the buyer.
What I ship
- JSON-LD schema graph. Organization, Article, BreadcrumbList, FAQ, Service, validated and monitored.
- Database-driven sitemap generator. Multi-index, filtered at the SQL layer.
- Canonical and redirect architecture in code. With tests.
- Core Web Vitals budgets in CI. LCP, INP, CLS — build fails on regression.
- AI-bot policy in robots.txt. Named per crawler, chosen per buyer.
- Search Console integration. Rich-result triage, structured-data regression alerts, post-deploy verification.
- Core Web Vitals dashboard. Chrome User Experience Report field data plus Lighthouse synthetic, on the buyer's URLs.
Where it fits
Replatform traffic drop
A site moved from one CMS to another. Organic traffic fell thirty percent inside sixty days. The cause is almost always mechanical — broken 301 map, lost canonicals, schema that no longer validates, or a sitemap pointing at pages the new platform serves as 404.
Rich results disappeared
The Search Console rich-results report turned red. Google tightened a validator and the JSON-LD that worked last quarter no longer parses. I read the actual output, map it back to the current Schema.org spec, fix it at the template layer, and verify clean validation across the next crawl cycle.
Page experience flagged as Poor
Search Console shows LCP and INP failing on mobile across a meaningful share of URLs. I pull the Chrome User Experience Report field data, isolate the offending routes, diagnose render-blocking resources or hydration cost, and ship the fix inside the build pipeline.
How I work
Every engagement opens with a written audit. Crawl-log sample, Search Console errors, schema output for the top page types, Core Web Vitals on real URLs, sitemap structure, redirect map, and the current AI-bot policy. The audit and a prioritized fix list ship before any code changes. Patterns from prior SEO engagements live in the research notes.
Engagement model
Audit-only engagements run two to three weeks and deliver a written report with a prioritized fix list and effort estimates. Audit-plus-remediation engagements run six to ten weeks, longer for replatforms or multi-locale sites. Retainer arrangements cover monthly Search Console triage, schema regression checks, and post-deploy verification for teams shipping continuously. To scope an audit, get in touch.