About
Danilson Ramos
Product engineer · SaaS, AI & full-stack developer · Zürich, Switzerland
Danilson Ramos is an independent product engineer based in Zürich, Switzerland. He designs, builds and operates production software — SaaS, custom software, AI-enabled products — in particular systems built around rules, workflows and payments, from product architecture and the data model to payments, email, deployment and observability. He is the creator of Atamatax, Horvera, Wabiro, Vellaci and SponsorsMetrics, and the author of the technical writing on bynilson.ch. Bynilson is the name of his practice.
Ten years writing software, five pairing it with design. Portuguese by heritage, French by birth, Swiss by address. I work alone, fixed fee, with founders and small companies across Switzerland and Europe — in English and French, with German and Italian for reading.
What I sell is not time: it is a first version that holds up in front of its first customer. The most honest proof I can give is the list of products I built and operate myself, with real payments.
In brief
- Name
- Danilson Ramos
- Role
- Independent product engineer — SaaS, AI and full-stack developer
- Where
- Zürich, Switzerland (Europe/Zurich) — remote across Switzerland and Europe
- Languages
- French (native) · English (fluent) · Portuguese (heritage) · German · Italian
- Stack
- Next.js · TypeScript · React · Postgres · Supabase · Stripe · Resend · Vercel · Python · LLM APIs
- Products
- Atamatax · Horvera · Wabiro · Vellaci · Neucelle · ComptaNexo · ScamHunter (in production) · SponsorsMetrics (private beta)
- Format
- Fixed fee, price agreed upfront — 1 to 2 projects per quarter
What I build
Six services, one method.
- Architecture and product auditTarget architecture, ranked risks, plan — before committing a build budget.
- SaaS productsAuth, roles, Postgres, Stripe, email, admin, deployment — the complete first version.
- AI-enabled productsExtraction, classification, structured generation — validated, evaluated, within budget.
- Next.js applicationsApp Router, Server Components, server actions, SEO and structured data built in.
- Custom web applicationsInternal tools, portals, rules engines, back offices for Swiss SMEs.
- SaaS V1What a first version must contain to survive its first customer.
Products
What I built and operate.
Each product is a case study: the context, the architecture decisions, what was difficult. No invented metrics.
- atamatax.comIn production · payments liveAtamatax — US expat tax infrastructureNext.jsTypeScriptPostgresStripeVercelRead the case study →
- horvera.comIn production · payments liveHorvera — flight compensation infrastructureNext.jsTypeScriptPostgresStripePostHogRead the case study →
- wabiro.comIn production · payments liveWabiro — identity and relationship tests built for search and sharingNext.jsTypeScriptPostgresStripeVercelRead the case study →
- vellaci.chIn production · subscriptions liveVellaci — EU Cyber Resilience Act compliance for product manufacturersNext.jsTypeScriptPostgresStripeCycloneDX · SPDXOSV · KEV · EPSSRead the case study →
- sponsorsmetrics.comPrivate beta · design partners onboardingSponsorsMetrics — the pricing cockpit for YouTube talent agentsNext.jsTypeScriptPostgresYouTube APIVercelRead the case study →
- neucelle.chIn production · subscriptions live · FR · DE · ITNeucelle — cash expenses and receipts for Swiss restaurantsNext.jsTypeScriptPostgresStripeVercelRead the case study →
- comptanexo.chIn production · free plan open · published pricingComptaNexo — Swiss pre-accounting, ready for the fiduciary hand-offNext.jsTypeScriptPostgresVercelRead the case study →
- scamhunter.chIn production · service operating · FR · ENScamHunter — intake and case preparation for a Swiss scam-investigation serviceNext.jsTypeScriptVercelRead the case study →
First-hand expertise
The topics I have scars from.
What I write comes from products I built, not from a summary of the documentation.
- 01Deterministic, versioned rules engines — eligibility (Horvera, EU261 / UK261 / CH), tax situation (Atamatax, PFIC / FBAR / FATCA), scoring (Wabiro)
- 02Production Next.js architecture — App Router, Server Components, server actions, caching, locale routing
- 03Stripe payments — checkout, subscriptions, signed webhooks, billing state in the database, fixed-price packages
- 04Search-driven acquisition — programmatic pages, structured data, multilingual sitemaps, content extractable by answer engines
- 05Third-party API integration — YouTube Data and Analytics in read-only OAuth (SponsorsMetrics), language-model providers
- 06Privacy by design — in-browser scoring (Wabiro), data scoped per account, Swiss FADP / GDPR compliance
- 07Cybersecurity — incident response, audits, authorised penetration testing; creator of ScamHunter.ch
- 08Product and pricing — fixed-price against percentage models, free result before payment, sharing loops
Writing
Technical notes and essays.
- SaaS MVP: build it with AI or hire? A worked exampleAn illustrative case, from a first orientation on three answers to an eight-stage matrix — and what would change the recommendation.SaaS engineering9 minRead →
- The limits of AI for building a web productWhat a coding agent does well, what needs evidence, and the minimum evidence to get — task by task, from the interface to maintenance.AI product engineering8 minRead →
- AI-generated website: why it looks generic, and how to fix itLovable, Bolt, v0 or Claude Code quickly deliver a page that holds together. What gives the AI away can be checked in the code: default titles, template text, preview links, a page that is empty without JavaScript. Here is what to look for, and in which order to fix it.AI product engineering6 minRead →
- My website looks good but doesn’t convert: ten causes you can checkDesign is often not the problem. Ten places where a visitor decides to leave, what you can check yourself, and how to validate a fix with little traffic.Building products10 minRead →
- What AI-generated SaaS founders should verify before launchAn app that renders is not a product ready to take money. Eight areas, one test per area, and the difference between declaring and verifying.SaaS engineering8 minRead →
- When should a SaaS use AI? And when it absolutely should notI have built three decision engines in production — tax, air passenger law, scoring — and none of them uses a language model to decide. Here is the rule I apply, the cases where AI belongs, and why RAG is often the wrong answer.AI product engineering8 minRead →
Working with me
How it goes.
A 20-minute scope call, free. A fixed-fee proposal within 48h — scope, deliverables, one price, dated milestones. A 4-to-6-week sprint with a demo every Friday on a preview. The full hand-off — repo, accounts, pipeline — and 30 days of replies within 24h.
I do not subcontract and I do not do hourly staffing. I take one to two projects per quarter to stay focused. If I am the wrong person, I say so on the first call.
A product to build?
One line is enough to start: what the product must do the day the first customer pays.
danilson.daveigaramos.pro@gmail.com · Reply within 24h, Mon–Fri.