Services
Four ways to work together
Each of these draws on the same underlying practice: understand the system before touching it, then build or explain the thing that holds up.
Full-Stack Development
Laravel/PHP and TypeScript/React, end to end
Backend and frontend built as one coherent system rather than handed off between specialists — Laravel/PHP APIs, TanStack-based React frontends, and PostgreSQL data models, with Bun-native services (Drizzle, better-auth, Biome) where a project calls for it.
- Laravel/PHP APIs and TanStack-based React frontends built end to end, from schema design through deployed UI
- Multi-tenant architecture and financial-grade data correctness for systems handling real transactions
- Comfortable across the stack: database design, background jobs, and the client-side data layer alike
AI Agent Architecture
Agent workflows built as production systems, not demos
Multi-step agent workflows with explicit state and observable history — treating a language model call as one step in a pipeline alongside ordinary application logic, not the architecture itself.
- Multi-step agent workflows with explicit state, not a single prompt wearing a chat interface
- Multi-channel delivery — web dashboards, Telegram, WhatsApp, Messenger — from one shared backend
- Versioned, auditable AI-driven changes, which matters anywhere an agent edits something a person is accountable for
Developer Relations
Making technical work legible to the people who need to evaluate it
Architecture blueprints and technical content written to be read and challenged rather than filed away — positioning technical work for whoever actually needs to understand it, whether that's a hiring manager, a stakeholder, or another engineer.
- Written architecture blueprints completed before implementation, sequencing decisions so they're cheap to challenge on paper
- Positioning technical work for the audience evaluating it, not just the engineers who'll maintain it
- Runs Crestion Career Consults, applying the same clarity to how professionals present their own work
Technical Writing
Long-form posts on the decisions behind the systems
Posts on backend architecture, AI agent engineering, and the TanStack frontend ecosystem, written from active work rather than assembled after the fact — favoring the trade-offs and reasoning over a feature list.
- Long-form technical posts documenting real decisions and trade-offs, not tutorials assembled after the fact
- Coverage spanning backend architecture, AI agent engineering, and the TanStack frontend ecosystem
- Comfortable writing for an engineering audience or a less technical one, depending on who needs to read it
Not sure which of these fits what you need?
Open to select engagements— describe what you're building and I'll tell you plainly whether it's a fit.
Start a conversation