, I'm Rodj.

I turn big ideas into engineered systems.

The Little Pond, Appledore · Childe Hassam, 1890

Works

Here are some of my works — production platforms and solo builds across AI, analytics, and product engineering.

Zalio

AI-native operations platform — a multi-tenant “company brain” with a governed LLM agent runtime.

Encore.dev · TypeScript · PostgreSQL · React

Case study

Klippy - Clipboard Manager

iOS and macOS Native Cross-Device Clipboard Manager

Case study Visit site

Gnosis Document Intelligence

Full-stack RAG platform with page-accurate citations and real-time streaming.

Next.js 16 · React 19 · Motia · Pinecone

Case study Visit site

Finance Analytics Dashboard

Multi-tenant SaaS that turns corporate card statements into AI-powered dashboards.

FastAPI · Next.js · Redis · ARQ

Case study

3Y Mode Growth Projections

Stock analysis platform with personalised three-year growth scenarios.

Next.js · Supabase · OpenAI · TypeScript

Case study Visit site

EnhanceMD Platform Refresh

Platform overhaul for a GLP-1 telehealth provider — design system, analytics, SEO.

Next.js · Design systems · PostHog · SEO

Case study Visit site

Reachify UI & SEO Revamp

Redesigned a restaurant phone-and-ordering platform for accessibility and local SEO.

Next.js · TailwindCSS · Accessibility · SEO

Case study Visit site

About

I’m a software engineer and AI automation enthusiast focused on AI-driven SaaS and internal platforms, RAG-powered analytics, and automation around finance and operations data.

I design and ship multi-tenant web systems that pair modern frontends (React, Next.js, TanStack Start) with scalable backends (Encore.dev, FastAPI, Node.js) and production-grade LLM integrations. I’ve led finance-analytics and credit-processing products that cut manual workflows by ~90%, while mentoring engineers and driving architecture across performance, scalability and security.

Experience

Built AI-native voice receptionist pipelines and human voice speed-to-lead pipelines for Australian businesses, ensuring rapid response times and multi-tenant architecture.
Architected scalable multi-tenant infrastructure supporting 24/7 human-operated and AI-driven lead response workflows with strict performance SLAs and response guarantees.
Integrated connected lead sources, custom SMS follow-up cadence flows, and real-time appointment booking/handover pipelines to maximize lead conversion rates.

Designed and built the foundational slice of a multi-tenant AI agent runtime (a “company brain”) on Encore.ts + PostgreSQL — an MCP server exposing governed, deny-by-default tools to LLM agents with citations, approval queues and full audit logging.
Hardened tenant isolation with PostgreSQL Row-Level Security (per-request SET LOCAL context, least-privilege role, fail-closed middleware) and pool tenancy modeled on the AWS Well-Architected SaaS Lens.
Implemented envelope-at-rest credential encryption — HashiCorp Vault Transit in production with an AES-256-GCM + HKDF fallback and zero-downtime key migration.
Built real-time, Messenger-class chat (WebSocket + NSQ Pub/Sub) and a cross-platform client (TanStack Start + Expo), deployed self-hosted via Docker / Coolify.