What if you got to build the technical foundation of the company redefining health and longevity?
About Helios
Helios is the human enhancement company. Almost all of medicine today is reactive, which means it waits for disease to appear and then tries to treat whatever shows up. We think the interesting problem is the opposite one: seeing trouble early enough to do something about it. That takes a kind of data the healthcare system has never collected, and we decided the only way to get it was to build the instrument ourselves. No one has built this before, and we think it is the largest opportunity in healthcare: the chance to build what medicine should have been all along.
That instrument is an at-home panel measuring over 1,000 biomarkers from a single draw, which makes it the highest-resolution at-home diagnostic available anywhere. The test is where we start, not where we stop. Around it we are building the brand, the platform, and the products that we expect will make Helios a household name in health, the company millions of people trust to live longer and better. Each draw becomes part of a longitudinal dataset connecting a member's biomarkers with their wearable data, their labs, their records, and how they actually live. We use that dataset to train the Helios Foundation Model (HFM), our path to medical superintelligence. Today, the HFM simulates physiology and predicts disease risk, and the HFM will soon design interventions outright. The first of those interventions go on sale this year, and every one we sell will send outcome data back into the model.
The three founders are Harvard Medical School physician-scientists (two of us helped pioneer CRISPR diagnostics) with backgrounds running from longevity research and computational biology through clinical practice and drug development. We are well-capitalized, and we are building a consumer company first. Roon (), our cognitive performance pouch, is the first of those products and is live now; the store, the funnel, and the systems behind it are part of what you would own. The products fund the data, the data compounds into models no competitor can train, and those models become our deepest moat. We are not improving the healthcare that exists. We are reinventing it, building a new kind of platform and the company that millions of people trust to monitor and run their health.
The model is the moat, and the model is only as good as the systems it lives in. You would be one of the first employees, and the only person whose full-time job is engineering. Everything technical runs through you: the models, the infrastructure they live on, the security that wraps them, and the product a member actually touches. In practice you would operate as our technical lead from the first day, and the role should grow into formal leadership as the team does.
We can cover the clinical and scientific side ourselves, since that is where the three of us have spent our careers. What we need is the person who turns the science into systems that survive contact with production. There is no lane to stay inside here, and whatever the company needs built next is yours to build.
What this role is not: a management job. For now there is no team to manage and no process to run, and the work is building the thing rather than overseeing the people who do. The seat can grow into technical leadership over time, but it starts hands-on and stays that way for a long while.
What you'll own
1. The website and member product
The core of the job is building the platform for the test: the website where someone learns what Helios is and orders a kit, and the member product on web and mobile where they connect their wearables and records and read what their results mean. That includes the report, the product's moment of truth, which has to be legible, honest about uncertainty, and beautiful enough that people screenshot it. You build the LLM agents that turn model outputs into language a person can act on, and the evaluation harnesses that keep them honest, because a fabricated number in a health report is the one failure we cannot tolerate. You make all of it something people want to open every week rather than a lab report they dread.
2. The backend
Behind the product sits the backend: the APIs, services, and integrations that take a member from checkout through lab-partner ingestion to results delivery. You design it, you build it, and you keep it up, with monitoring, alerting, and observability deep enough to chase a problem from a model output all the way down to a frontend bug, because members make health decisions based on what we show them.
3. Data security
We hold sensitive health data belonging to real people, so security is a design constraint from the first commit rather than a compliance exercise bolted on at the end. Encryption, access control, audit logging, and secrets management all run through you, our vendors get held to the same standard, and the whole system operates at HIPAA-grade requirements. When something goes wrong anyway, you run the response, ideally on an architecture you designed so that no single failure can expose very much.
4. Serving models at scale
Research code has to become production inference, and what serves a thousand members today needs to serve a hundred thousand without a rewrite. Latency and reliability will live in your head constantly, and so will cost, since inference bills have a way of growing faster than anyone planned for.
5. Data and training infrastructure
You own the machinery that makes the models possible: the pipelines from lab-partner ingestion through QC and normalization, datasets versioned carefully enough that a result from six months ago can still be reproduced, and the training and experiment infrastructure that keeps all of it maintainable as new data sources come online, wearables and records first, our own genomic and metabolomic assays later.
6. Training the models
And then the models themselves, trained on data from our biomarker panel. Validation deserves special mention, because when a model's output feeds someone's health decisions, leakage and miscalibration stop being academic concerns. As new data sources come online, the new models are yours to train as well.
Who you are
We are hiring one extraordinary engineer, not filling a role. The bar is simply the best person we can find. Show us what you have shipped and owned, because that will count for more than where you worked or went to school. Specifically, we are looking for:
Track record. You have built systems end to end that real users depended on. You can point to something real: a product in production, a model you trained and operated, a system that survived scale. Proof counts for more than pedigree
A spike. Unusually deep in at least one of ML, backend and infrastructure, security, or product engineering, and hungry to cover the rest
High learning rate. You pick up unfamiliar domains fast, learn from what production tells you, and change your approach quickly when something is not working
Bias to action. You ship systems, not design docs. You do not wait for permission to build the prototype
Low ego. The bar is high, but no important problem is beneath us: model training in the morning, a React bug in the afternoon, an incident runbook at night
Unusually good at something. Maybe it is a side project people actually use, a company of your own, or an open-source tool with real users. We hire for spikes, not résumés
Four or more years building production software, though exceptional work overrides the number
A degree in computer science or a closely related technical field, or equivalent depth you taught yourself. Real, demonstrable experience matters more than the credential, but we expect the fundamentals
Real production ML experience, meaning you have trained models and operated them yourself, and you know the unglamorous parts like leakage checks and calibration on high-dimensional data
Strong Python and backend fundamentals, including API design and cloud infrastructure on AWS or GCP
Hands-on LLM engineering, from prompt design through agent architecture and evaluation
Enough TypeScript and React to ship real interfaces without help
Hands-on data-security experience with sensitive or regulated data, including encryption, access control, secrets management, and secure architecture. You treat real people's health data as if it were your own. This is a hard requirement, not a nice-to-have.
A workflow already built around the current generation of AI tools, used to raise your own ceiling rather than lower the bar. The work that leaves your hands should be better for how you used them, not noisier. You own the judgment, the taste, and the final result
The self-direction to take a vague problem to a shipped solution, and the judgment to pick the right next thing without being told
Energized by ambiguity rather than slowed by it, comfortable defining what good looks like before anyone else has
Bonus points
Omics or bioinformatics experience, or comfort with high-dimensional biological data
Healthcare data standards or compliance work such as FHIR or HIPAA
MLOps and observability tooling for production ML systems
Founding-engineer or early-startup experience, including a company of your own
Vector databases and retrieval systems
Open-source work or public technical writing we can read
Shipped a consumer product people chose to use, not had to use
Strong opinions, loosely held, on how AI changes how software gets built
Who thrives here
The people who do well in a company this small bias toward action and finish what they start, and they would rather learn from real users than from a planning cycle. They see the whole machine without losing the details, and they bring their own ideas instead of waiting for a spec, then argue for them from evidence, with us most of all. They delegate to AI tools the way a good manager delegates to a team, aggressively but with verification, because that is the only way one engineer covers this much ground. They treat member data the way they would want their own treated, and they would sooner miss a deadline than cut a corner on it. Mostly, they want to build something enormous, and they have the patience to spend years doing it. Above all, they build large, complex, yet maintainable systems in a fraction of the time most strong engineers think possible, and they would rather work next to a few people that good than a hundred who are not. The best of them simply get more done than seems possible, taking a system from idea to production in days, doing a team's worth of work alone, and shipping a quarter's worth of roadmap in a week. They run hot and move fast, and they expect the early years of something this ambitious to ask a great deal of them. They have a high learning rate and high agency: they would rather ship the prototype than schedule the meeting, and they treat every result, good or bad, as fuel for the next one.
Why this role
The honest case for this role over a bigger title or a bigger paycheck somewhere else comes down to ownership. The entire technical foundation of the company would be yours, with the authority to make the calls and the obligation to live with the results. You would work directly with the three founders, with no layers between you and any decision that matters. Your code would reach users within your first few weeks, and if the company becomes what we intend it to become, the systems underneath it will be the ones you built. Your compensation is a competitive salary plus meaningful equity, so you share in what you help build.
Details
Commitment: Full-time
Location: Boston preferred. We would love to have you in person with the founders and can help cover relocation, but we are open to remote for the right person.
Compensation: Competitive salary plus meaningful equity, so you share in what you help build
The interview
The process is simple. You meet each of the three founders. If we both like what we see, we bring you in for a paid one-week work trial: building something real with us, not a whiteboard exercise. If it clicks, we make you an offer.
How to apply
Submit whatever you are proudest of building, whether that is a repo or a product running in production. What was it? What made it hard? What did it produce? How did you use AI tools while building it? We care less about polish and more about proof that you can build. Send your CV or résumé along with it. The work tells us the most, but we want to see the full picture.
The bar is high and we will hold it, but do not screen yourself out. If you are not sure you meet every line above and you suspect you might be exceptional, write to us anyway and let us be the judge.
Helios is an equal-opportunity employer. We do not discriminate on the basis of any characteristic protected by applicable law, and we welcome applicants from every background.
Source: Y Combinator job 103939. Listed on YC when retrieved 2026-11-02.
