Founding Scientist – Peptide Chemistry
Listed on 2026-06-17
-
Software Development
Machine Learning/ ML Engineer
Founding Scientist (Peptide Chemistry)
San Francisco
· Full-time · $120–180K depending on what you bring
$10k referral bonus for successful hires (half equity, half cash).
It's Monday, 10 AM. You walk through the glass doors at the San Francisco lab, grab a coffee, grab a Red Bull, and pull up the overnight dashboard.
Yesterday's cell and mouse data has already been crunched. The house AI parsed the PK curves, flagged the outliers, and ranked the candidates while you slept. Two peptides from last week's batch showed decent half-lives. One was a dud. You skim the summary, make a note, and move on.
At noon, the head of science from a major AI lab walks in. He's got access to a frontier model that isn't public yet and wants to understand your assay data well enough to train on it. You spend forty-five minutes at the whiteboard explaining dose-response curves to someone who understands transformers better than anyone alive but has never pipetted anything.
At 3 PM, the last step is done. Synthesized, purified, QC'd. That peptide didn't exist this morning. You made it.
You hand it off to the in vitro team. At 4 PM it's tested in cells. Most do roughly what you'd expect — modest activity, nothing to write home about.
But one of them is oddly potent. Ten-fold over the nearest analog. We checked it twice. You may have just made the most potent peptide of its class ever synthesized.
In the next room, an automation engineer finishes programming the Open Trons to run our assays faster. She calls you over to validate the protocol under your supervision — by next week, one step takes half the time.
By 5 PM it's dosed in mice. By 8 PM you have preliminary data.
One day. Synthesis to animal data. One day.
The team goes to dinner at the poké spot in Mission Bay. Someone argues about whether the potency data is real or a lucky artifact. You think it's real, but you don't say that yet. You'll know tomorrow.
The RoleYou'll be one of the founding scientists at the company. We're targeting first-in-human dosing by mid-2026. If we hit it, this is also a world record for moving a non-vaccine therapeutic into the clinic from scratch.
Most biotechs take weeks to go from synthesis to animal data. Here, it happens in one day. You synthesize peptides in the morning, run them through our cell assays by afternoon, test top candidates in mice by evening, and hand the results to our ML team overnight to design the next version. We do not know of another lab on earth running this loop at this speed.
Start-ups are unique. Roles aren't as ossified or constrained as in a large corporation— you’ll often learn a new role every few months, enabling you to grow as a person and team member. Keep this space for growth and variability in mind when reading this rough outline of what working together could look like.
As we are spinning out multiple deep research projects, if you want to lead your own program, we are open to that. Weird ideas are welcome!
More concretely, you will…
- Run Fmoc SPPS on the Liberty Blue — design the sequences, own the synthesis, troubleshoot failed couplings and difficult aggregating peptides
- Purify on the Agilent 1260 and 1290 prep HPLC — method development, fraction collection, yield optimization
- Confirm structure and purity by LCMS on the Agilent 6530 — you read a spectrum and know immediately if something is wrong
- Hand candidates to the in vitro team with a clear synthesis summary and QC data they can trust
- Pick up in vivo work over time — mouse dosing, PK sampling, plasma processing — prior experience not required, good hands and curiosity are
- Troubleshoot on the fly — failed coupling, mystery yield loss, peptide that won't dissolve — you don't wait to be told what to try next
What you'll get
- If we hit our 2026 timeline, your name is on the fastest non-vaccine drug to reach the clinic.
- In a year you'll have run synthesis, cell assays, animal studies, and ML-driven design.
- More than most scientists have learned in a decade.
- Your experiments directly determine what goes into humans.
You might be a fit if:
- You picked up a hard technical skill (flow cytometry, mass spec, animal work) in weeks, not semesters
- You are a deeptech…
(If this job is in fact in your jurisdiction, then you may be using a Proxy or VPN to access this site, and to progress further, you should change your connectivity to another mobile device or PC).