
InkBorne
AI-powered story and comic generation product. Built on the Anthropic API alongside ComfyUI and Ollama, in a production Django, Celery and Redis pipeline. Pre-launch.
Available for select project engagements with SMEs, government and private clients. Twenty years of production software, with extensive hands-on experience in agentic AI.

AI-powered story and comic generation product. Built on the Anthropic API alongside ComfyUI and Ollama, in a production Django, Celery and Redis pipeline. Pre-launch.
Async multiplayer space trading sim set in a living, seasonal galaxy. Unity 6 WebGL client with a Node and TypeScript backend (Fastify, Prisma, PostgreSQL, Redis), and LLM-driven narrative generation. In active development.

A toolkit for the Firefly III personal-finance platform. Built in Django with HTMX and D3.js. Functional, and in weekly use for my family finances.

Hands-on track in agentic LLM and RAG self-feedback loops, focused on developer productivity and code-implementation tooling. Directly informs how I design AI integrations for client engagements.
I am Don Bloomfield, a Sydney-based senior software and systems consultant with twenty years of production work behind me. My background spans full-stack web, data and ETL pipelines, cloud architecture, simulation and VR, with extensive recent experience in production GenAI tooling. I lead engagements with diagnosis: finding the actual cause of a problem before deciding what to build. I hold RHCSA and ITIL v4 Foundations, and am currently studying ITIL v5 Strategic Leader.
I work directly with clients. The person you brief is the person who builds it, and what we agree on is what you get.
Project engagements typically run two to six months, end to end. I take on a small number at a time. Shorter advisory or discovery engagements available as a starting point, or as a sanity check on a brief before a build.
Production GenAI pipelines (Anthropic API, Ollama, ComfyUI), agentic and RAG architectures, and AI tooling that earns its place inside a business. Cloud architecture and managed hosting on Azure and AWS, with the governance scaffolding to keep it safe.
Django, Python and SQL systems for SMEs and government clients. Database design, ETL, REST APIs, full-stack web. Built to last past handover, with documentation and reliability instincts that come from twenty years of running production systems.
Linux administration, COTS / GOTS integration, identity and access governance, automation and root-cause troubleshooting. Senior systems and IT operations work for high-availability environments.
VR / XR, simulation (DIS / HLA / C4I), Unity production, broadcast and live-data systems. Specialist work taken on selectively where the brief fits.
I spend time early in an engagement understanding what's actually happening, before deciding what to build. Twenty years of production systems work has shown me how much downstream time that early effort tends to save.
Vague briefs waste money. I take time to scope properly before quoting: what success looks like, what the constraints are, what the right shape of delivery is. Once we agree, you have a clear plan and one accountable person seeing it through.
Documentation, reliability instincts and operational discipline are part of every delivery, not a separate workstream. If I leave and you hire someone else to maintain the system, they should be able to pick it up without ringing me.
Where AI accelerates real work, I build with it. Where it adds risk or opacity without value, I say so. The judgement comes from having shipped both kinds of work in production.
When the ACT government rolled out the MyWay public-transport system, Council on the Ageing's existing client database needed to interface with the new transit data on an ongoing basis. I designed and delivered the API integration and database services across two major project revisions, ensuring continuity of the seniors-card workflow while the new transport platform was rolled out.
Result: Integration running in production through two platform revisions; client retained for ongoing services.
Contracted as systems and integration lead for an international high-availability ADF training delivery system out of HQJOC, I traced a long-standing reliability and throughput issue to root cause and redesigned the affected stage of the ingest pipeline.
Result: Roughly 25% of the ingest workload eliminated. Validated under live exercise conditions in 2024.
Designing and shipping a production GenAI pipeline for an AI-powered story and comic-generation product. Anthropic API for narrative generation, ComfyUI for image generation, Ollama for local-model fallbacks, deployed on Django with Celery and Redis. Demonstrates orchestration, prompt and context engineering, and production deployment of LLM workflows.
Result: Pre-launch product, currently in active development.
The fastest way to start a conversation is the form below. Replies within 24 hours.