In 2024, Paul Conyngham's eight-year-old rescue dog Rosie — a Staffordshire bull terrier he'd adopted from a Sydney shelter in 2019 — was diagnosed with advanced mast cell cancer. Large tumors had appeared on one of her back legs. Surgery helped. Chemotherapy slowed the spread. But nothing could shrink what was already there.
Veterinarians gave Rosie months to live.
Conyngham, a Sydney-based tech entrepreneur and data engineer with 17 years of machine learning experience, accepted neither the prognosis nor the helplessness that came with it. He did what any good engineer does when confronted with an impossible problem.
He started asking better questions.
Conyngham's first call wasn't to a specialist. It was to a chatbot.
He fed Rosie's health information into ChatGPT. The AI recommended immunotherapy — and, remarkably, directed him to a specific institution: the University of New South Wales Ramaciotti Centre for Genomics. He reached out. They agreed to help.
For $3,000, Conyngham paid to have two DNA sequences run: one from Rosie's healthy blood, one from her tumor. The comparison would reveal exactly where her cancer lived — the precise mutations driving it.
"The idea is you take the healthy DNA out of her blood and then you take the DNA out of her tumor and you sequence both of them to see exactly where the mutations have occurred," Conyngham told Australia's Today program.
With that data in hand, he turned to AlphaFold — Google DeepMind's protein structure prediction tool — to model the mutated proteins in three dimensions. He identified a driver protein called c-KIT, central to mast cell cancer in dogs, and pinpointed drug targets.
The output: a half-page mRNA sequence formula. No biology degree required.
Conyngham and his collaborators at UNSW — now calling themselves "Team Rosie" — identified an immunotherapy drug that looked promising. They contacted the pharmaceutical company that made it. They requested compassionate use.
The company refused.
Another team might have stopped there. Conyngham pivoted.
Martin Smith, director of the UNSW Ramaciotti Centre for Genomics, suggested a different path: build a custom mRNA vaccine. The same platform technology that powered Pfizer-BioNTech and Moderna's COVID vaccines — lipid nanoparticles carrying mRNA blueprints — could, in theory, be turned into a personalized cancer treatment.
Conyngham, using Grok to refine the final design, handed his formula to Professor Páll Thordarson, director of the UNSW RNA Institute. Thordarson's team synthesized the physical vaccine.
From formula to finished vaccine: less than two months.
There was one remaining obstacle: ethics approval. UNSW researchers couldn't administer an experimental vaccine without it. Enter Professor Rachel Allavena at the University of Queensland, a canine immunotherapy specialist who already held the necessary approvals for experimental treatments.
In December 2025, Conyngham drove ten hours from Sydney to Gatton, Queensland. Rosie received her first injection.
One week later, the tumor began visibly shrinking.
By January 2026, Rosie had enough energy to jump a fence chasing a rabbit — a dog who, just weeks earlier, had been too lethargic to move.
"At the start of December her mobility was way down, she started to shut down and be a bit sad. And towards the end of January, she was jumping over a fence to chase a rabbit."
— Paul Conyngham
The tennis ball-sized tumor on her leg had shrunk by roughly 75%.
Researchers at UNSW confirmed what Conyngham had quietly suspected: this is the first personalized cancer vaccine ever designed for a dog.
The implications weren't lost on anyone involved.
"It raises the question, if we can do this for a dog, why aren't we rolling this out to all humans with cancer?"
— Martin Smith, Director, UNSW Ramaciotti Centre for Genomics
Professor Thordarson was equally direct about what Rosie's story reveals:
"What Rosie is teaching us is that personalized medicine can be very effective, and done in a time-sensitive manner, with mRNA technology."
He also noted something more profound: Conyngham's ability to generate a viable mRNA recipe with zero biology training demonstrates that AI is democratizing the entire process of cancer vaccine design.
Scientists are careful to note what this was: one dog, one tumor, zero controlled trials. Mast cell tumors can occasionally shrink on their own. A second tumor on Rosie did not respond to the first vaccine. Conyngham is already designing a second treatment to target it.
This is not a cure. It is not a clinical trial. It is a proof of concept — and an extraordinary one.
The same mRNA platform technology Conyngham accessed for under $3,000 is currently moving through human oncology pipelines with hundreds of millions of dollars behind it. Moderna and Merck's jointly developed personalized melanoma vaccine has shown a 49% reduction in cancer recurrence or death in Phase IIb trials. Phase III trials are now underway for melanoma and non-small-cell lung cancer.
The science is real. The timeline is compressing. And a rescue dog from Sydney just helped prove it.
Paul Conyngham is not a symbol of DIY medicine run amok. He is a preview of what personalized cancer treatment looks like when the barriers — cost, time, institutional gatekeeping — are systematically removed by technology.
For decades, personalized cancer vaccines were theoretically possible but practically inaccessible: too slow to design, too expensive to manufacture, too complex for anyone without a lab. Rosie's case compresses that timeline to two monthsand that cost to $3,000.
For comparison, Moderna and Merck's version, when approved for human patients, is expected to cost between $100,000 and $300,000 per patient.
Professor Thordarson believes the approach can be democratized in Australia without dependence on foreign pharmaceutical companies. He also sees applications that extend beyond cancer — including neurological diseases.
Conyngham is now working with everyone involved to explore how other dog owners might access the same process. He is designing a second vaccine for Rosie. And the team that helped build it is thinking much bigger.
One rescue dog. One data engineer. One freely available chatbot. A $3,000 DNA sequencing bill. And a world-first in personalized medicine that has university scientists stunned and asking questions no one in pharma has adequately answered.
Rosie isn't cured. But she's chasing rabbits again.
And the technology that got her there is coming for human cancer next.
Original research and reporting by UNSW Sydney. Story published March 2026.
