jmnicholson

Scite (scite.ai) | Senior Developer | REMOTE (US)

We build AI-powered tools that help researchers and organizations make sense of the scientific literature. Small team, hard problems, real momentum. Our MCP server just launched, enterprise adoption is growing, and the product roadmap is full of ambitious bets.

Stack:

Frontend: React 18, Redux/Redux Saga, Webpack, SWC, Express (SSR), CSS Modules Backend: FastAPI (Python 3.11), Flask, PostgreSQL, Elasticsearch, Redis, Celery AI/ML: Claude (Anthropic), OpenAI, PyTorch, Hugging Face Transformers, sentence-transformers, RAG pipelines

Infrastructure: AWS (ECS Fargate, RDS, S3, Lambda), Docker, Kong API Gateway, GitHub Actions, CircleCI

Observability: OpenTelemetry, New Relic, Sentry, PostHog

Other: Stripe, SendGrid, Playwright, TypeScript

What we're looking for:

Strong engineering fundamentals Comfort working across the stack, Python and JavaScript/React in particular Experience with PostgreSQL, Elasticsearch, or similar data-intensive systems is a plus

Familiarity with LLM APIs and retrieval-augmented generation (RAG) is a bonus Genuine curiosity about how AI can improve how people interact with research

If you're interested or know someone who might be a great fit, email josh@scite.ai

I'm a co-founder at Scite. We've built an MCP server that lets AI assistants search and read full-text scientific articles.

We have 50M+ full-text articles from licensed publisher agreements. If your institution subscribes to a journal, you can authenticate with your institutional credentials and your AI assistant gets the same access you'd have through your library. No scraping, no workarounds, fully licensed.

For anything outside your subscription, you get metadata, abstracts, and direct linkouts to the publisher page.

It works with Claude, Cursor, and anything that supports MCP.

Scite.ai | Full-Stack Software Engineers | Remote (US/Canada)| Full time | https://scite.ai/

We are hiring for multiple engineering roles at scite.ai to build the worlds best place to search, discover, understand, and access research.

We have unique access to the world's literature through publisher partnerships and and have developed a market leading generative AI Assistant.

We are mostly looking for full stack senior software engineering roles who have experience and interest in scaling up systems and building large scale ingestion and search - based applications. This role requires no previous AI experience or knowledge (though its a plus!) so its perfect for folks who want to break into the generative AI space from an engineering background.

Senior Full Stack Engineer (Remote, US/Canada)

https://researchsol.bamboohr.com/careers/49?source=aWQ9MzU%3...

Senior UI/UX Designer (Remote, US/Canada) https://researchsol.bamboohr.com/careers/50?source=aWQ9MzU%3...

The results in the link are literally how hacker news has been mentioned in scientific papers. Not sure how you consider that clickbait or misleading or why you consider those meaningless.

More here too: https://scite.ai/search/citations?page=1&q=%22pardon%20the%2...

Meta (a great name) was a very promising startup a few years back in scholarly publishing. It had successfully analyzed millions of articles to help with discovery of research. It's a shame it is being shuttered but there are a lot of tools out there now doing similar things and in many cases much much more:

scite.ai (I am co-founder) semantic scholar

Yes, there is a lot more that goes into it. The details of our indexing are described in detail in a recent paper of ours: https://direct.mit.edu/qss/article/doi/10.1162/qss_a_00146/1...

Is this on Citation Statement search? I think it is probably because the citation context contains two or more citations in it. We look at citations per sentence so it is duplicated there. I can see how that is confusing though.

We can't release the full dataset as our licensing agreements with publishers restrict it. We do have an API though that can be used: https://api.scite.ai/docs

check out our work in this area: https://scite.ai/search/citations

We have extracted 918M citation statements from 27M full text articles.

We (scite.ai) have extracted 918M citation statements (3 full sentences) from 27M full-text articles (more than half are from paywalled articles from indexing agreements).

We've taken this an extra step and have classified the citation statements as providing supporting or contrasting evidence (https://direct.mit.edu/qss/article/doi/10.1162/qss_a_00146/1...)

And, recently, have made these citation statements easily searchable to find expert analyses and opinions on nearly any topic extracted from the literature: https://scite.ai/search/citations

We're effectively bringing shepardizing to science at scite (scite.ai). Here's a piece I published on the topic recently: https://www.sciencedirect.com/science/article/pii/S259023852...

We're trying to fix this problem at scite.ai https://medium.com/scite/scite-trust-but-verify-7d13e6a144cc

We're effectively creating the rotten tomatoes for science at scite.ai. We rely upon a deep learning model to classify citation statements as supporting, contradicting, or mentioning.

https://www.statnews.com/2018/12/13/solving-fake-news-proble...

How is this any better than using the sci-hub plugin?

Agree! It's really ridiculous and quite sad. I am no longer at a University so basically everything I read is stolen via Scihub.

Hopefully, more and more researcher will continue to preprint their work. Something we're trying to help with at https://www.authorea.com (disclosure: I work there)

Super admirable goal and super tough task. Kudos on the progress!

Kind of crazy that arXiv doesn't make this easier to do itself. We wrote a piece on how a future arXiv might look which seems relevant:

The arXiv of the future will not look like the arXiv https://www.authorea.com/users/3/articles/173764-the-arxiv-o...

Interesting side story is that Gadjusek did a year in prison for child molestation (which he think is a natural/good thing).

There is a documentary on him on BBC called the Genius and the boys https://www.youtube.com/watch?v=4OxppDxzSww

Authorea fully supports LaTeX and MathML with a WYSIWYG interface like GDocs, Paper, or Word.

https://www.authorea.com/product

Is this the anti-Kardashian index? https://genomebiology.biomedcentral.com/articles/10.1186/s13...

I think science is at an interesting point. We've just had our Napster moment with Sci-hub. We'll soon have our Youtube moments with Authorea (https://www.youtube.com/watch?v=Qa1ObxI_dqU) and the rise of preprints in new disciplines, In short, I see self-publishing and dissemination increasing in the coming years.

Maybe we'll have our Justin Bieber of research quite soon!

You should try writing these on authorea.com (get DOI and permanent archival). Basically, turning them into official pieces of open post-publication peer review. Could cross-post to your blog too still!

We're working to provide a Word-like interface on git repositories for researchers to write their manuscripts in at authorea.com. Our editor allows researchers to write in markdown, LaTeX, or rich text all within one article.

both! I went to UC Santa Cruz.

This is a great point. We need to provide researchers with options that don't kill their career prospects. One idea, which we're working on at Authorea (authorea.com) is to help researchers write their documents online so they can be easily published as preprints. This way open access can be achieved in a manner that goes beyond a PDF or Word doc, and researchers can still play the game so long as it exists.

Some more info here: https://www.authorea.com/123679/d8cErAb_olPC5CvcWWwsig

exactly!

the idea is here: https://thewinnower.com/papers/the-r-factor-a-measure-of-sci...

Not a homework question. I am developing a new metric for reproducibility in scholarly publishing and trying to understand previous examples of successful models.