A Man With No Biology Degree Used ChatGPT to Build a Cancer Vaccine for His Dying Dog

Welcome back to SavvyMonk, your go-to source for AI and tech news that actually matters.

Today's story might be the most remarkable thing I've covered in this newsletter. A man in Sydney used ChatGPT and Google's AlphaFold to design a custom cancer vaccine for his dog. The tumor shrank by 75%. Scientists are stunned. And the implications for human medicine are enormous.

Let's get into it.

Hate writing docs? Same.

Mintlify built something clever: swap "github.com" with "mintlify.com" in any public repo URL and get a fully structured, branded documentation site.

Under the hood, AI agents study your codebase before writing a single word. They scrape your README, pull brand colors, analyze your API surface, and build a structural plan first. The result? Docs that actually make sense, not the rambling, contradictory mess most AI generators spit out.

Parallel subagents then write each section simultaneously, slashing generation time nearly in half. A final validation sweep catches broken links and loose ends before you ever see it.

What used to take weeks of painful blank-page staring is now a few minutes of editing something that already exists.

Try it on any open-source project you love. You might be surprised how close to ready it already is.

Try Atlas today!

Paul Conyngham is a Sydney-based machine learning engineer. He co-founded Core Intelligence Technologies, a data science consultancy, and served as a director of the Data Science and AI Association of Australia. He has 17 years of experience in AI and data analysis. He does not have a background in biology or medicine.

In 2019, Conyngham adopted Rosie, a Staffordshire Bull Terrier-Shar Pei cross, from a shelter. She'd been abandoned in bushland. He calls her his best mate.

In 2024, large tumors appeared on one of Rosie's back legs. The diagnosis was mast cell cancer, the most common skin cancer in dogs and notoriously challenging to treat once it spreads. Conyngham spent tens of thousands of dollars on chemotherapy and surgery. The treatments slowed the tumors. They did not shrink them. Vets gave Rosie one to six months to live.

The AI-Powered Plan

This is where it gets wild.

Conyngham opened ChatGPT and started asking what else was possible. The chatbot suggested immunotherapy as a treatment direction and pointed him toward genomic sequencing at UNSW's Ramaciotti Centre for Genomics. He paid $3,000 AUD to get Rosie's DNA sequenced, comparing healthy cells from her blood against DNA from her tumor to identify the exact mutations driving the cancer.

Then he went further. He fed that data through AlphaFold, Google DeepMind's protein structure prediction tool (which won the Nobel Prize in Chemistry in 2024), to model the 3D shapes of proteins encoded by Rosie's mutated genes.

Using his machine learning algorithms, he identified which mutated proteins, called neoantigens, were most likely to trigger an immune response.

The output was a half-page formula describing the mRNA sequence for a vaccine targeting Rosie's specific cancer mutations.

Martin Smith, associate professor of computational biology and director of the Ramaciotti Centre, was initially skeptical. But Conyngham's analysis was solid. Smith later told The Australian that Conyngham was “relentless” and that his work left the genomics team stunned.

The Vaccine

Conyngham first tried to get an existing immunotherapy drug for Rosie, but the manufacturer refused to provide it for compassionate use in a dog. That's when Smith connected him with Pall Thordarson, director of the UNSW RNA Institute.

Thordarson, an Icelandic nanomedicine expert, used Conyngham's data to create a custom mRNA vaccine. The same technology behind the Pfizer and Moderna COVID-19 vaccines. From sequence to finished vaccine: less than two months.

But Conyngham still needed someone with ethics approval to administer it. Mari Maeda of the Canine Cancer Alliance connected him with Rachel Allavena, a professor of canine immunotherapy at the University of Queensland, who had the required approvals. Conyngham drove 10 hours with Rosie for her first injection in December 2025, returned for a booster in January 2026, with another planned for March.

This was, according to Thordarson, the first personalized cancer vaccine ever designed for a dog.

The Results

Within one month, the tennis ball-sized tumor on Rosie's hock had shrunk by roughly 75%. Her energy returned. Six weeks after treatment, she spotted a rabbit at the dog park and jumped a fence to chase it.

Smith's reaction: “It was like, holy crap, it worked!”

Not everything responded. Some tumors did not shrink. Conyngham is now working on a second vaccine targeted at the resistant tumor, sequencing it to understand why it did not respond. He's been clear that this is not a cure. But it has bought Rosie significantly more time and a better quality of life.

Why This Matters

This story is not really about one dog. It's about what becomes possible when powerful AI tools are freely available to people who know how to use them.

Here's the context: Moderna and Merck are running billion-dollar Phase 3 trials on mRNA-4157 (V940), a personalized cancer vaccine for humans that works on the same principle. Sequence a patient's tumor, identify mutations, build a custom vaccine. Their five-year data shows it cut melanoma recurrence by 49% when combined with immunotherapy. Expected cost per patient when approved: $100,000 to $300,000. Expected approval: around 2027. Over 120 similar clinical trials are currently running worldwide.

Conyngham did it for tens of thousands of dollars with free AI tools and university lab access.

Now, there are important caveats. Veterinary experimental treatments face much lighter regulatory scrutiny than human medicine. There's no veterinary equivalent of FDA Phase I-III trials for a one-off compassionate use case. ChatGPT didn't design the vaccine by itself. It served as a research assistant and planning tool. The actual vaccine design, synthesis, and administration required real scientists with deep expertise. And this is a single case, not a clinical trial. One dog's response does not prove the approach works broadly.

But the signal is impossible to ignore. The tools to design personalized medicine already exist. They're available to anyone with an internet connection and the skills to use them. The bottleneck is no longer the science. It's the regulatory and economic systems built for a world where designing a treatment took a decade, not eight weeks.

Thordarson put it directly: “What Rosie is teaching us is that personalized medicine can be very effective and done in a time-sensitive manner, with mRNA technology.”

The Bottom Line

A data scientist with zero biology training used ChatGPT and AlphaFold to design a personalized cancer vaccine that shrank his dog's tumor by 75%. The same approach is the basis of billion-dollar pharmaceutical programs that won't reach patients for years. The technology is ready. The infrastructure is not. Rosie's story is a proof of concept for a future where AI makes precision medicine fast, accessible, and personal. That future is closer than most people realize.

Category: Research Planning

“I have [describe your problem or goal]. I have no formal background in [relevant field]. Act as a research consultant and help me build a step-by-step plan to investigate this problem. For each step, explain what tools or resources I'd need, what skills are required, and where I might find collaborators or experts. Flag any steps where professional oversight is essential.”

What sticks with me about this story isn't the technology. It's the fact that Conyngham spent three months writing a 100-page ethics application just to treat his dog. The hardest part of building a cancer vaccine wasn't the science. It was the paperwork. That tells you something about where the real bottlenecks are.

Hit reply; I read every response.

See you in the next one.

— Vivek

P.S. Know someone who works in biotech or AI or just loves a good dog story? Forward this their way. They can subscribe at https://savvymonk.beehiiv.com/