I build reliable, practical software and ship useful products. I focus on readable code, clear architecture, and measurable outcomes. When I'm not building software I assemble and repair custom, high-performance PCs.
- 🔭 Currently: building tools and dashboards to automate workflows and improve business productivity.
- 🌱 Learning: scalable backend patterns, observability, and developer tooling.
- 👯 Open to: collaboration on open-source projects and automation tooling.
- ⚡ Fun fact: I enjoy turning difficult edge cases into simple, well-testable solutions.
-
Frontend / Full-Stack
Next.js (App Router + React Server Components + Server Actions)
React / React Native (mobile & web)
Tailwind CSS + shadcn/ui + Radix UI (clean, accessible, customizable UI) -
AI & LLM Integration
Vercel AI SDK (multi-provider: ,routers.)
Prompt engineering + RAG patterns (via templates & custom implementations)
n8n (AI agents, workflows, enterprise integrations) -
Backend & Data
Node.js / TypeScript APIs
Drizzle ORM + Neon Serverless Postgres
Vercel Blob (file handling)
FastAPI (Python APIs — in planning / active use) -
Authentication & Security
Auth.js (NextAuth) — OAuth, JWT, secure sessions
Offline-first & secure patterns (mobile safety focus)
I design systems for clarity, and operational simplicity:
- Start with a modular monolith using Clean Architecture (entities → use cases → adapters). Decompose into microservices only when warranted by scale or separation of concerns.
- Prefer explicit, well-documented interfaces (REST/GraphQL) and small, well-scoped services.
- Make observability first-class: structured logs, metrics, traces, and health checks.
- Security-first approach: authenticated APIs (OAuth2 / JWT), encryption in transit and at rest, and minimal data retention.
- Practical scalability: caching, background workers, async queues, and horizontal autoscaling where needed.
AI & automation specifics:
- Use performant inference stacks (TGI / vLLM) with cost/latency trade-offs in mind.
- Combine vector stores + hybrid search for RAG workflows.
- Coordinate multi-step automations with lightweight orchestrators or workflow tools (n8n, task queues).
- Default to safe data handling: anonymization, rate limits, conservative telemetry.
(links to public repos; private repos a)
- autto-marketer- — Scrapes Google for marketing data and generates marketing assets via a GUI and web interface. (JavaScript)
- n8n-templates--borrowed-from-a-sail-25-private-account — Collection of n8n automation templates and enterprise workflows; integrations for Gmail, WhatsApp, and common services.
- gpt-engineer-sm444 — CLI platform for experimenting with programmatic code generation; precursor to https://lovable.dev. (Python, MIT)
- whatsapp-MCP — Message-sending service for WhatsApp with media support. (Python, MIT)
- dyad-for-SC — Fork of dyad-sh/dyad; a local, open-source app builder and developer tooling. (TypeScript, Apache-2.0)
- auto-scribe-buddy — (private) Backend that uses Notion as a content source to discover trends and generate blog drafts.
- Bloglandingpagedesign — Notion-backed blog landing page generator. (TypeScript)
- THe-leap-project- — Dashboard and admin UI for the Leap project. (TypeScript)
- kovera-prime — (private) TypeScript project in active development.
- graph-sitter-1 — Scriptable interface to a multi-lingual language server (fork). (Python, Apache-2.0)
- Linda — Mobile app focused on safety and routing for Kenyan users; anonymity and offline-first features.
- nextjs-ai-chatbot — (private) Next.js chatbot project (TypeScript).
- Small, fast iterations with automated tests and CI keeping main branch deployable.
- Clear PRs and focused reviews.
- Prioritize simplicity and maintainability over premature optimization.
If you want contact details or a portfolio link added, tell me which to include:
- Email: gfordevowrks@gmail.com -Website: www.sailcraftsolutions.co.ke
Made with focus for — © 2026