PostEra is a Y Combinator-backed company using machine learning, applied to medicinal chemistry, in order to expedite the discovery of new drugs and get new cures to patients faster.
What are we working on?
Partnerships: We partner with drug hunters and help them get cures to patients faster. We're currently collaborating with Pfizer and several small biotechs, using all of PostEra's ML-powered software to help their drug discovery.
Manifold: The first-of-its-kind SaaS platform for scientists to get molecules made.
COVID Moonshot: PostEra is helping lead the world's largest open-source drug discovery effort. Moonshot is an international team of scientists working on an antiviral cure for COVID and in 1 year we've gone from ideas on a website to potent compounds which should soon be ready for human trials preparation.
Machine Learning Researcher Boston, MA, United States or Remote
Senior Software Engineer Boston, MA, United States or Remote
Why join us?
Mission: Use technology to find cures for diseases!
Traction: PostEra has secured a multi-year ML collaboration with Pfizer while also leading the world's largest open-science initiative to find a COVID cure.
Finances: We raised a $2.5m seed round coming out of Y Combinator from tech titans like Sam Altman. Even better, PostEra is now profitable.
Engineering at PostEra
Engineers at PostEra work closely with the product team and drug discovery scientists to ensure they are building high impact tools for drug discovery. Meetings within the engineering team can be more frequent, but each engineer presents their work to other engineers, scientists, and the product team every two weeks. Engineers are also encouraged to become involved where they see fit in product and drug discovery strategy discussions. Code is tracked and reviewed through Github, and we attempt to ensure quick turnaround on Pull Requests.
Chemists or machines often design molecules that would be interesting as potential therapies for certain diseases. However, making these molecules is not always straightforward. Historically,
recipes for these molecules are multistep processes involving chemical reactions designed entirely by humans. However, we build tools to better design these complicated recipes. This involves searching over permutations of thousands of reactions and billions of building blocks, which quickly grows in computational complexity. Thus, we have to engineer systems that effectively search and prune synthesis trees reminiscent of the ever expanding game trees in Chess or Go.
We also frequently have to search for interesting molecules similar to a given query molecule or drug of interest. This search can be over billions of similar vectors or graphs and requires state-of-the-art vector distance and subgraph searching approaches.
Our current ability to search for similar molecules in large databases is limited to databases of ~1 billion molecules, and very specific search types defined by 3rd party software. The speed, expense, and lack of flexibility of this existing solution is becoming troublesome. With the advances in nearest neighbor vector searching brought about by machine learning, we are looking to improve on our existing solution and improve our infrastructure to easily be able to search ~10 Billion molecules with different, more flexible search methods.
When a chemist goes to perform a reaction in a lab, they often want to look for precedence in the existing patent or academic literature. From that paper, they could then see how high yielding the reaction is if run with certain chemicals present. We currently provide useful prediction based off millions of patents; however, we need better ways for chemists to interactively find the exact documents and reactions they want. Improving this searching interface for chemists to find useful precedence, would help our scientists speed up our drug discovery efforts.
Working at PostEra
Medicines come before models. We're all nerds at heart, either for science, ML or engineering but ultimately the mission of getting new cures to patients is more important than the technology we use. As such, everyone is very focused on innovation only so far as it helps our mission.
We go out of our way to find and interview talent from non-traditional backgrounds. The quality of candidate is never compromised but we want new people who interact with PostEra to be able to look at the team and say
hey, here's someone I can gel with.
As part of our commitment to an inclusive workplace, we are happy to offer prospective engineers the chance to connect with our engineering employees who come from underrepresented backgrounds. It’s a way to get a better sense of our team and what it might be like to work with us.
If you’re interested in connecting with our team, be sure to bring this up during one of our introductory calls!
Work from Home
Interested in this company?
Skip straight to final-round interviews by applying through Triplebyte.