We are a PhD-led engineering team building AI systems for biotech, healthcare, and life sciences. With 20+ PhD experts and over a decade of development experience, we combine research depth with practical engineering. Our work spans drug discovery, omics data, diagnostics, and AI infrastructure. We value clarity, scientific rigor, and collaboration.
We learn from each other. We share knowledge. We build things that matter. If you want your work to contribute to real research, we’d love to meet you.
Scientific Content Lead
Turn real scientific work into the content that makes an AI engineering company credible
About the role
Blackthorn AI builds production AI and data systems for organisations working with scientific, clinical and sensor data. Our team publishes in Nature and presents at NeurIPS. Our engineers hold PhDs in the domains they build for.
Almost none of that is visible on the internet.
We are hiring one person to fix that – not by writing marketing copy, but by producing original technical work and publishing it. Benchmarks we actually ran. Architectures we actually built. Comparisons nobody else has bothered to do properly. The kind of material a Head of Computational Biology reads and thinks these people know what they are doing.
This is a research job with a distribution problem attached, which is why we are hiring a scientist rather than a marketer. We will teach you the marketing.
What you will do
Produce original technical artefacts (about half your time). Run and publish reproducible benchmarks. Real examples from our current plan:
- Protein structure prediction compared across AlphaFold3, Boltz-2, OpenFold3 and Chai-1 on a fixed held-out target set, re-run quarterly
- Nextflow vs Snakemake vs WDL on the same whole-genome workload, same hardware, with cost and wall-clock published
- Vector RAG vs GraphRAG on a held-out PubMed question set, with the evaluation set released
- Differential expression method selection compared across DESeq2, edgeR and limma on three GEO datasets
Every piece we publish must contain something that did not exist before you made it. If a competent person with ChatGPT could produce it in an afternoon, we do not publish it.
Write for buyers, not reviewers (about a quarter of your time). Take that work and write it so a VP of Engineering or a Head of Data at a medtech or biotech company understands it in five minutes and trusts it. This translation is the hardest part of the job and the reason most academic content fails commercially.
Run the content operation (about a quarter of your time). Track SEO/AIEO trends. Own the content map and publishing calendar. Interview our senior scientists and turn a 30-minute call into a publishable piece under their byline. Implement pages, schema markup and internal linking. Track what happens afterwards. Refine hypotheses and the plan if needed.
How success is measured
- Citation in AI answers. How often Blackthorn is named when someone asks ChatGPT, Claude, Gemini or Perplexity a question in our domain.
- Website traffic – number of people visited the Blackthorn website.
- Positions of the pages in search engines.
- Original artefacts shipped. Benchmarks, datasets, reference architectures.
- Sales and partner usability. Whether our founders and partners actually send your pages to prospects.
- Pipeline nurture. Our deals take six to eighteen months to close. Whether the people we met at a conference are still hearing from us in month nine.
Must have
- Self-motivated, with genuine interest in biomedicine, AI, and in how technical work reaches an audience
- Fast learner. We will teach you SEO, AI search optimisation, positioning and content operations from scratch
- Problem solver. You will often be the only person on this, and you will have to decide things without asking
- Degree or ongoing research in AI, machine learning, data science, computer science, or life/natural sciences
- Able to run an experiment end to end – design it, code it, get a result, and defend the method when someone challenges it
- Write clearly in English for a technical reader. Similar to but not exactly academic writing
- Python scripting and comfort in a terminal. Enough to run pipelines, wrangle data, produce figures, and script against an API
- Willing to say no to topics. Half this job is deciding what not to publish. Publishing broadly damages us
Nice to have
- Peer-reviewed publications
- Literature and marketing research
- Understanding of SEO, AI search optimisation (AEO/GEO), or topical authority
- Experience with systematic literature review or market research
- Customer or user research experience
- WordPress, JSON-LD schema, Google Search Console, sitemaps
- Video: turning a technical result into a five-minute explainer
- Domain background in genomics, structural biology, clinical data, or computer vision
What you get
- A named byline on everything you produce, and co-authorship where the work merits it
- Direct mentorship from our CEO and access to our CSO, principal scientists and scientific advisors
- Compute and tool budget for the benchmarks
- A portfolio of published technical work that is useful whether you stay in industry or return to academia
- Remote, flexible hours, compatible with a PhD or postdoc schedule
- Part-time from roughly 20 hours per week, with a route to full-time
To apply
Send us:
- One thing you have written or built that you are proud of, and two sentences on why
- Your answer to this: pick any two open-source tools in a domain you know. Which is better and how would you prove it? Half a page is plenty
- A CV or a LinkedIn link