Vitaly Kuzenkov
Technical Delivery Lead & AI Solutions
A technical delivery lead and the founder of VistFlow. I spent a decade shipping software for blue chip clients (CBA, Telstra, Mastercard, NSW Health, Fanatics, TAB) on projects up to A$2M, and I now build production AI myself. That means I close the gap between what a business needs and what gets built, and I can talk to the stakeholders and read the code in the same meeting. Honest reporting and clean delivery are how I work.
“Curiosity, shipped.”
A decade of delivery, then building it myself.
Every phase deeper into shipping, every phase more ownership: from delivering enterprise software for blue chip clients, to designing, building and running production AI on infrastructure I own.
Founder and AI Solution Builder
VistFlow · Sydney, Australia (remote)
Founder of VistFlow, building production AI applications and automation for businesses, from the first idea to deployed and running, describing the work in plain language while AI writes the code. I bring a decade of enterprise delivery discipline to my own AI builds.
- Shipped a production Customer Success Intelligence platform for a high-value client: it answers plain-English questions over their live business data and scores client health daily, thousands of customer interactions a month.
- Architect and run my own personal cloud, 150 containers across 70 stacks, private by default, with nightly backups and everything as code.
- Design, build, secure and run my own AI products: a trading assistant, a job hunter, and a native macOS Claude workspace.
Senior Agile Delivery Manager
Aura · Sydney, Australia
Led agile delivery for engineering teams building software for Fanatics and TAB, owning planning, delivery cadence and stakeholder communication end to end.
- Delivered software engagements for clients including Fanatics and TAB.
- Ran the full agile cadence across cross functional teams: planning, standups, reviews, and retrospectives.
- Kept delivery and client expectations aligned through clear, honest reporting.
Senior Digital Producer
Imagination · Sydney, Australia
Produced large experiential and technology projects (roughly A$150K to A$2M each) for blue chip clients at a global experiential agency. This is where I learned to ship big, polished, client facing work on time.
- Delivered flagship work for CBA, Telstra, Mastercard, AWS, and the City of Sydney.
- Produced builds such as the Telstra Insight Ring, the Mastercard Experience Centre, and the CBA Innovation Lab.
- Owned budgets, vendors, and timelines on projects from A$150K to A$2M.
IT Project Manager
PicNet and Metallurgical Systems · Sydney, Australia
Managed IT and software delivery for public sector and industrial clients, the start of a decade in Australian tech delivery.
- Delivered projects for NSW Health and Guide Dogs NSW.
- Coordinated requirements, vendors, and rollout across technical and non technical stakeholders.
Distinction, and the proof of it.
A Bachelor of Science in Information Technology from UTS, with Distinction, plus the academic honours that came with it.
Bachelor of Science, Information Technology, Enterprise Systems Development
Master's Degree, Economics, Project Management in the Petroleum Industry
- Information Technology Undergraduate Achievement Prize, awarded for the highest weighted average mark.
- Dean's List 2015, top 10 undergraduate IT students.
Evidence, not a portfolio.
The proof I take an idea from nothing to live, secured and operated, end to end. Production AI I built, and the blue chip experiences I delivered before that.
Customer Success Intelligence platform
AI agents that answer questions about a business in plain language.
A production, multi-tenant platform I built for a high-value client. Two specialist Claude agents reason over their live delivery and marketing data: they rank what is at risk, surface what shipped, and answer plain-English questions with the SQL and citations shown underneath. A daily command center turns every client into one health score with a clear next action. It replaced about ninety minutes of manual reporting each morning, so the team spends its time deciding what to do rather than assembling the picture.
- Two specialist Claude agents reasoning over live operational and marketing data.
- A daily health score per client, with the SQL and citations visible under every answer.
- Organisation-based row-level security, encrypted per-client keys, multi-provider AI failover.
Telstra Insight Ring
A flagship interactive brand experience.
A flagship interactive experience delivered for Telstra: large, client facing, shipped on time and on budget during my decade in Australian enterprise delivery.
Mastercard Experience Centre
An experiential build for a global brand.
An experience centre delivered for Mastercard: a large, polished, client facing build I owned from budget to handover.
CBA Innovation Lab
Interactive delivery for Commonwealth Bank.
An innovation lab delivered for Commonwealth Bank: hands-on delivery of a polished, interactive experience for a blue chip client.
Cortex
A native macOS Claude workspace I built and use daily.
A native Mac app I built and use every day. It drives the Claude Agent SDK as its engine: multi-agent chat, a file tree, a live context meter, session restore, and a read-only diff of every edit. It is token and origin gated, after I had Claude find and close a security hole in it. This is the tool I build everything else with.
- Drives the Claude Agent SDK rather than reimplementing the agent.
- Multiple concurrent agents, session restore, and a live context meter.
- Runs on my Claude subscription, not metered API.
My Cloud Brain
My self-hosted cloud, 150 containers across 70 stacks, that I own and run.
150 containers across 70 stacks on a server I own and run: my own AI products, an embeddings service, a vector and SQL data layer, plus my photos, files and mail. Private by default over Tailscale, nightly encrypted backups, everything as code. This is the substrate the rest of my work runs on, and proof I can stand up and operate real systems alone.
- 150 containers across 70 stacks, private by default over Tailscale, fronted by Caddy and Cloudflare.
- Nightly encrypted backups and fully infrastructure as code.
- My own AI products, my data, my mail, all on hardware I control.
Hands
A computer-use agent that sees my screen and acts for me.
A native macOS agent I built that can see the screen and act on it: it drives Chrome directly and can also use the operating system itself, keyboard and mouse. I teach it a task once by recording it, then it replays it. Everything it may touch is fenced by an explicit permission model that fails closed, with a full audit of every run.
- Two independent lanes: a browser lane driving Chrome, and an operating-system lane.
- Fail-closed permissions: it can do nothing that has not been explicitly granted.
- Record a task once, replay it; every run is audited.
How I work.
If this resonates, let's talk. I would rather show you than tell you.
vitalykuzenkov@gmail.com- LinkedInwww.linkedin.com/in/kuzenkov
- CVcv.vitalyroo.com
- Upworkupwork.com/freelancers/kuzenkov
- Portfoliocv.vitalyroo.com/portfolio
- Based inSydney, Australia






