Damian Yazbeck.

AI Engineer · Available now

4 years in tech, 2.5 in AI. Full-stack agentic AI platform spanning 49 operational areas - built, deployed, and maintained by one person.

AI Engineer. Production-first. Sydney, AU.

One engineer, 9 months solo. 40+ containers, 280+ workflows, 10+ production bots, a fine-tuned 30B model, 100,000+ lines of architecture docs - all running 24/7 on a dedicated server.

See my work Get in touch
0
n8n workflows
0
containers
0
production bots
0
operational areas
0
lines of arch docs
$0.00
30B fine-tune
Selected work

Everything I've shipped & what I'm shipping.

Live products, an open-source framework and bot platform, and what's on the bench right now. Flip Plain / Technical up top to switch every card between recruiter-readable and engineer-deep.

Status
Domain
Under the hood

Architecture, not screenshots.

Real data-flow diagrams for the systems behind the products. Click any diagram to magnify it, or open the project for the full breakdown.

Technical skills

Stack at a glance.

What I actually use in production - not a buzzword list.

AI & Automation

Claude / Anthropic API OpenAI Agents SDK Mistral Groq Ollama Agentic AI Multi-agent systems Prompt engineering RAG QLoRA fine-tuning n8n MCP

Backend & APIs

Python (async) aiogram Pydantic V2 FastAPI REST APIs Webhooks Twilio Vapi Cal.com Baserow PostgreSQL Redis

Infrastructure

Docker / Compose Linux / Ubuntu Nginx Systemd Hetzner bare-metal SSL / TLS Git / CI-CD Qdrant Tailscale VPN

LLMOps & Observability

Langfuse Grafana Uptime Kuma GoAccess Automated deployment pipelines Production traces
How I build

Three disciplines, applied every time.

Not philosophy. Rules I actually break builds against.

Scope, then execute

Every build starts with one hard question: does this solve a real problem for a real person, or just my urge to build? Once it clears that bar I commit and ship; when it does not, I say so. ScopeShield shipped, found no traction, and got cut in weeks; Bridge stays a pet project because that is the honest label. Scope it straight, then execute without second-guessing.

Write it before you build it

The 100,000+ lines of architecture docs are not a record of what got built - they are how I work out what to build. If I cannot write a clear sentence about why a component exists and what contract it holds, I have not understood the problem yet. The docs force the thinking; the thinking prevents the rework.

Boring infrastructure first

Observability, routing, and failure handling come before features, always. Every Meridian bot launched with Langfuse traces on day one. Synqr's voice stack sat gated while the proxy layer, data residency, and key security were proven. The interesting work is only safe once the boring work is done.

Background

Production-proven.

4 years in tech. 2.5 years in AI. Before that, years managing complex operations across retail and fitness - the shift is what I build with, not how I think.

AI Engineer · DamianBuilds (Oxide / Atlas Method) 280+ workflows, 10+ bots, 40+ containers, 49 operational areas running continuously. Open-source lean methodology for Claude Code (Atlas Method v1.1.3).
Automotive AI QA · Thundersoft Speech recognition QA and tool-call evaluation for in-car AI systems. Tested whether the model triggered the right action - opening windows, dialling correctly, calling the right tool - and captured behaviour bugs across the AI/vehicle integration surface. Effectively LLM evaluation work before evals had a job title.
Driver Platform Support · Uber Driver-facing platform support. Reproduced app and trip bugs in real test environments, escalated through Jira to engineering. Triage between ground-truth and the platform team.
Product Support Specialist · Glofox Customer-facing SaaS support for gym/studio software. Reproduced ~5 bugs per day in dummy environments, ticketed through to engineering. Effectively manual QA feeding the eng pipeline.

I've also led teams in past roles - 8+ staff across two gyms as a Gym Manager, multi-property operations as a Building Manager. Comfortable in cross-functional work and managing complexity.

Hire the operator, not the demo.

I ship agentic AI in production, solo. Imagine what I do on a team. Open to AI Engineer roles in Sydney or remote across Australia. Available immediately.

damian@damianbuilds.com