Engineering leadership · Melbourne

Rohyl Dwaine.

I co-founded Block Zero, a bootstrapped trading platform, and built its engineering team as CTO.
At NewDoc, I’ve been building the platform and AI agents.

I’m interested in how teams work with AI: how they ship, make decisions, and spend less time repeating themselves.

Experience

2020—2026

Block Zero

Jun 2022
Jul 2026

Co-founder & CTO

Led engineering at a bootstrapped trading company, with engineering leads reporting to me. Ran hiring, performance reviews and technical direction.

14engineers
at peak

Four years building the company and its engineering team.

Engineering leadership

11 direct reports

We grew from 5 to 14 engineers in six months. I split us into a product team and a platform team, gave every service one owner with its own dashboard, SLOs and on-call, and moved from writing code to reviewing about 1 in 5 of everyone else’s pull requests.

CTO

Frontend
Backend
Infra
Data

Decisions in writing

255 RFCs in 20 months

I pioneered an internal RFC culture, resulting in 255 RFCs across 17 authors in 20 months, 31 of them mine.

Production agent platform

Built the initial Rust MCP server for agent access to the trading platform, with scoped, revocable permissions.

AI agents

Scoped access

Trading tools

Users can revoke access

One API, many tools

One OpenAPI spec, built from our Rust routes, generated the TypeScript SDK, the Rust SDK, the CLI, the docs and a 60-tool MCP server, with a check that flags any endpoint missing from an SDK. I set the architecture and the team built it.

Languages

Rust
TypeScript

Data

ClickHouse
Redis
RisingWave

Messaging

Kafka
NATS
gRPC
Soketi

AI

MCP
Anthropic Managed Agents

Platform

Kubernetes
GitOps
Own servers in Amsterdam
Tokio

Development

Claude Code

Agent-assisted development and reusable release workflows.

Team size is peak engineering headcount. Direct-report count spans my time managing the team.

NewDoc

Apr–Sep
2026

Founding engineer

Built systems for NewDoc’s operations team and its first AI agent, with staff review before use.

4operational systems
in production

Selected work from my first five months.

Patient messaging

A$20k saved per month

Moved patient messaging off Zapier and into NewDoc’s admin workflow, cutting ongoing vendor costs while giving staff delivery history and consistent opt-out handling.

Telephony & SMS

3 → 1 patient phone numbers

Moved voice and SMS onto NewDoc’s own Twilio account and consolidated the numbers patients used to contact the team.

3 phone numbers

One patient numberVoice + SMS · Twilio

Doctor pay runs

95 doctors

Replaced spreadsheet pay calculations with an invoice-preparation workflow for staff review.

Spreadsheets

Draft invoicesReviewed by staff

Roster management

1,000+ shifts managed

Gave staff one place to manage shift requests and apply them to the clinical schedule.

An AI agent for clinics

Reads patient messages and drafts replies for staff to approve. Every claim in a draft has to cite something the model was shown, or staff see nothing. Health data stays in Australia: if anything would leave, the agent refuses to run.