I'm Natan Valencia. 21 years in IT, working solo from the first line of code to deploy: backend, data, frontend and infrastructure. Everything below I built myself, and it's all running.
The same pattern in every project: sources with no API, data that doesn't line up, and at the end an app or a report that answers one concrete question.
I designed, built and deployed both of them on my own. The same kind of work I deliver in an MVP.
SEO auditing with my own crawler: 84 detectors across 10 categories —cannibalization, indexing, hreflang, JS rendering, regressions between crawls—, Google Search Console integration, a rank tracker with SERP data and report exports to Excel and Word. Python, FastAPI and PostgreSQL with raw SQL: no ORM, no JS framework.
The main source is a web app with no public API, so the system intercepts around twenty internal APIs per product and has extracted close to 50,000 of them, unified with the Spanish funds from the CNMV and the US ones from the SEC: three sources with incompatible formats and update frequencies turned into one comparable model. On top of that, a faceted search over millions of rows, portfolio backtesting, a comparison tool and a fund graveyard with pre-death patterns.
You have the idea and you need to show it working — to users, to a partner, to an investor. I build it the same way I build my own: from the first line to deploy, with every decision documented.
We put in writing what's in and what's out. Fixed price means fixed scope: whatever comes up along the way goes on the v2 list instead of blowing up the deadline.
Development on the same stack that runs my own products (FastAPI + PostgreSQL, Astro or Svelte). From week 2 there's a real URL where you can watch it progress — a demo every week, no PowerPoints.
In production, on your domain. The repository is yours from day one, with the decision log documented so any developer can carry on without me.
Strategy development to order and validation before risking a single euro: thousands of backtests per phase on top of a queryable results database, and deliverables generated automatically from the data — every number in the report can be traced back to the run that produced it.
The client for the last study came along after seeing FatCow's public reports. That's why this site shows deliverables instead of telling you how good I am:
It monitors the P&L of trading accounts in real time against the Rithmic API and, when an account crosses its loss limit, cancels the orders and flattens the positions on its own. Multi-tenant on top of an API designed for a single session, a circuit breaker for gateway outages and 24/7 protection that doesn't depend on keeping a browser open.
ApeSpider's crawl backend: it crawls concurrently and resumably —it survives a container restart mid-job—, tells a real page apart from a WAF block that returns a 200, decides on its own when it needs to render JavaScript, and delivers results through signed callbacks. No frontend: pure backend, 547 tests.
Most small websites run legal texts copied from another site, when the law requires them to reflect what your site actually does. Two products on the same engine: a web app with a step-by-step wizard and PDF export, and a WordPress plugin that scans the real site —active plugins plus analysis of the public HTML—, infers which services affect compliance (GDPR, LSSI-CE, AI Act) and publishes the legal pages straight into WordPress. The plugin is finished and in legal review ahead of launch.
My SEO blog and, along the way, a technical case study: a full migration from WordPress to Astro with a custom parser for the SQL dump —63 articles converted—, 290 301/410 redirects without losing a single URL with history, and Open Graph images generated automatically on every build.
I started out administering servers —networks, DNS, Windows and Linux domains for an organization with seven sites— and that foundation still runs the show: logs, performance and infrastructure have never been "another department's problem".
Along the way: 13 years of technical SEO under my own brand, four of them as the in-house specialist for an international portal of more than 20,000 URLs in 7 languages, working in English every day. Today I build my own products and systems to order. That double role defines how I work: I understand the business behind the data and I build the software that exploits it.
usual stack: Python (FastAPI · PostgreSQL · Celery) · Sentry · Astro on Netlify (and headless WordPress) · Docker on VPS
Every project comes with its decision log documented. If a number shows up in a report of mine, it can be traced back to the query that generated it.
A scraper nobody can get working for you, data you have but don't exploit, a bot you want to validate before risking money, a backend that's falling short. Write to me and I'll tell you straight whether I can help, how, and what it would cost. And if I can't, I'll tell you that too.