Product engineer

You can’t walk past a crooked picture frame.

We are a funded team of two with three customers, a working app, a map of 7.5 million US healthcare providers, and room for an engineer. Clinical trials run late because the physicians who see the right patients never hear about the study, so we reach those physicians directly. In one trial, 89 of the 107 patients identified were interested.

You will own the product and the map, and sit next to the people using them.

Nothing here is ever really finished, because every conversation with a user changes what comes next. If you want a spec, this is the wrong job. If you want to engage us about what to build, read on.

Four in five clinical trials miss their enrollment timeline. The patients exist. The physicians who see them are twenty minutes from the site and have never heard of the study.

Spellbinder closes that gap. We mapped 7.5 million US healthcare providers, resolved from 400+ sources into one profile each: what they treat, who they share patients with, where those patients live and which trials they have enrolled into.

Two agents work the map. One reads a protocol and ranks physicians for every site, every day. The other reaches them by email and phone, then hands them to the coordinator when they say yes. You click a button and a patient shows up at the site.

It works. In one central nervous system trial, still ramping, 107 patients were identified and 89 were interested and sent to the site. That is 83%, because the ask came from their own physician and not an ad.

The same map now serves biopharma commercial teams and healthtech companies selling to providers. We are two people. You would be our only engineer.

AI is revolutionizing how drugs are discovered. The way they are tested and brought to market is still archaic, and trials still lean on marketing agencies and ads to find patients. We want people who will rethink it from first principles, with care.

We are building software that is almost invisible. We do not measure it by stickiness or time in the app. The work gets done on your behalf. We measure ourselves in the wild: are patients arriving at sites, are we finding the right physicians, and are physicians engaged? Shipping the feature is only the start.

The goal is bigger than one product. We are building an entire distribution system for healthcare, with physicians fully engaged.

This is healthcare. How we treat patients sits at the center of everything we build, alongside how we treat our users.

We want an engineer who is also a product person. You will build quickly, put it in front of users, find out what is wrong and build it again.

You will work directly with the founders, with nothing in between. You will be in the room when we decide what to build, who to build it for and where the company goes next.

This is not a role where someone hands you a spec. You will learn what a coordinator really does with the morning list by sitting next to her while she uses it. Then you will change the product, often the same week.

It is three jobs, and they feed each other.

Build the product. You are walking into a working app and healthcare map, and it is the first version of what could be best in class software. Take it there. Make the architecture calls, ship across the stack and set the standards the next engineers inherit.

Keep the map true. The healthcare map is the company. You will maintain and extend the pipelines behind it: new sources, better entity resolution, better predictions and the monthly refresh. When a physician tells us a profile is wrong, you will find out why.

Build in the seams. Sit in the room with site teams, sponsors, commercial teams and providers. The best product ideas here live between their workflows, where work gets dropped, copied by hand or chased by phone. Fix it for one customer, then make it part of the product for all of them.

• Week 1: Ship to production. Sit in on real customer calls.

• Day 30: You have spent time in person with a site team and a sponsor. You have told us which of our assumptions are wrong.

• Day 60: You own one core surface end to end, from what it should do to how it runs in production.

• Day 90: The foundation is in place: deploys, monitoring and tests where they matter. We can measure map quality against what physicians and coordinators tell us. Something you shipped has moved a real number: physicians engaged or patients at a site. You have a point of view on what we build next, and we are arguing about it.

Picture a meeting with us. Amr comes from product design and architecture, and James from product, finance and biochemistry, so we rarely see a problem the same way. We will disagree, egos stay out of it, and we leave with a solution.

Someone will crack a joke about startups or the industry. Then we get back to work.

This is not a stuffy environment, and there is no fake startup theater. We are real people. We do not drink the Kool-Aid, or any other aid, and we are not high on our own supply.

We are relentless about the goal and open minded about the path. If the evidence says pivot, we pivot. We trust each other to put in the work, and we will trust you the same way. What you put into this company will also shape everyone who joins after you.

We love testing anything that makes us faster or helps us build more reliably.

When a new tool or way of working beats what we do today, we change our process and put it in place.

The care we put into our users is the level we expect from everybody here. They are used to marketing agencies and ads, and we want to be a breath of fresh air: professional and, above all, human. When we reach out to a physician, we adapt our voice to that physician.

• Opinionated. You will tell us when we are wrong and show your reasoning. When a user proves you wrong, you change your mind without drama.

• Product oriented. You care what you build and whether it worked for the person using it. You have taste, and you notice when software is careless.

• Hardworking and fast. You test the smallest real version and kill what does not work. Every delay is a patient waiting, and we will be straight about what this job takes.

• Kind. You will be in rooms with busy coordinators and physicians who owe us nothing. You listen well, and you make the people around you better.

• 5+ years building and shipping full-stack products, ideally with real time at an early-stage company

• Designed systems from scratch and lived with the consequences

• Worked with large, messy data: pipelines, data modeling and matching records that do not want to match

• Put AI features or agents into production, with a view on evals and failure modes

• Worked directly with customers, and enjoyed it

• Use AI coding tools as a normal part of how you work

Nice to have, not required: entity resolution at scale, voice agents, forward deployed or founder experience. If you have never worked in healthcare, we will teach you. We need you to be curious about it.

We would rather you hear it now.

• It is early. There is no product manager, no team and no process beyond what we build together.

• The app and the map work, and they are first versions. Parts of them will need rethinking as we grow, and the map is never finished.

• Not everything is automated yet. You will help decide what changes first.

• Some of what you build will be thrown away within a month. That is the point, and it can still sting.

• $150K to $190K base and 1% to 2% equity, on a 4-year vest with a 1-year cliff

• Health, dental and vision coverage

• A straight answer on funding and runway in our first conversation

• Founders who have built at Palantir, Squarespace, Amazon, N26 and Remesh

• Conversation with Amr and James. Show us something you built, and tell us what you got wrong.

• Technical deep dive. A system you designed, and the tradeoffs you would remake.

• References, then an offer with the full numbers explained.

Email hello@spellbinder.ai with something you built and care about, what you learned from its users and one thing you would change about Spellbinder. No cover letter needed.

Spellbinder is an equal opportunity employer. We consider all qualified applicants without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, veteran status or any other protected status.