Palm | Devpost

Inspiration

Healthcare's attention ( and its money) flows toward the end of life. Longevity is the buzzword of the decade: supplements, biohacking, "healthspan" , "biological age" for adults who already carry decades of accumulated habits and damage. We think the market is looking in the wrong direction.## Inspiration

Health doesn't start at 40. It starts at conception. The first 1,000 days of life - from conception to a child's second birthday - are a critical window in which nutrition and environment leave lifelong epigenetic marks. This is the DOHaD framework (Developmental Origins of Health and Disease), tracing back to David Barker's work linking early-life conditions to adult cardiovascular and metabolic disease. The science is uncontroversial: a healthy baby is the precondition for a healthy adult. And the first 1,000 days produce the highest density of health data of any life stage — weight, growth percentiles, feeding, sleep, milestones, vaccines. The data is being generated. It's just being lost.

This stopped being an abstraction when one of our co-founders gave birth to her second child , who was admitted to the NICU at Bambin Gesu Hospital in Rome, and spent several weeks there. By any measure an exceptionally capable person, she found herself completely lost, not in the medicine, but in managing it. The most precious thing in front of her, the health of her child, and she couldn't get a grip on the information around it.

The information existed! THAT was the maddening part. A discharge PDF here. A photo of a paper report there. A scrap of paper where a night-shift nurse had scribbled a comment that turned out to matter. WhatsApp threads, email attachments. It was all there, just scattered across formats and channels, impossible to assemble into a coherent picture of her own child.

That's where PALM started: in the gap between "the data exists" and "I can actually use it when it counts."

What it does

PALM brings healthcare back into the palm of your hands. For pediatrics, that means two connected products: a mobile app for parents, and a clinical dashboard for the pediatrician. A single place where a child's health actually lives. It rests on three pillars:

It KNOWS your child. PALM centralizes a child's medical history into one structured, searchable source: discharge letters, lab results, growth curves, vaccines, prescriptions, the photo of the paper report. It ingests any format- a snapshot of a handwritten note, a PDF of a digital exam- and turns it into catalogued, retrievable, understandable information. The scattered scraps become one record.

It LEARNS your child. Every question asked, symptom logged, and document uploaded makes the picture richer. A fever reported for a healthy five-year-old and a fever reported for a premature infant in post-discharge follow-up are not the same question- and PALM weighs each one against that specific child's history.

It SUPPORTS you. When a parent has a doubt at 2 a.m., PALM answers from what it knows about their child and from a RAG anchored to validated pediatric protocols and clinical guidelines (SIP, SIN, hospital protocols), not the open internet. And because a parent's hands are often full- literally holding the baby - we're building real-time voice support: a hands-free conversational assistant (OpenAI's realtime model, with ElevenLabs tuned for medical terminology) that walks a caregiver through the moment as it happens, like cleaning a newborn's umbilical stump or checking a feed, speaking the answer aloud instead of making them stop and read.

The other side: a dashboard that gives pediatricians their time back

PALM isn't only for parents. The companion dashboard turns the messy reality of a child's history into something a doctor can read in seconds.

Take a baby discharged from the NICU, one of the most data-dense, hardest-to-track situations in all of medicine. From a single discharge letter, the dashboard structures everything that matters: gestational age and corrected age (so every growth and development milestone is measured correctly, not against the chronological calendar), birth and discharge auxology with percentiles, the respiratory and feeding profile, prior morbidities, home therapies, and the schedule of protected follow-ups. Then it layers on what the family logs at home day by day- weights, feeds, diapers, sleep, medications, symptoms like reflux or irritability.

The result: instead of reconstructing a fragmented history during a seven-minute visit, the pediatrician opens PALM to a pre-assembled, longitudinal clinical picture, and spends the visit on care, not detective work.

Critically, PALM does not diagnose and does not replace the pediatrician. It is an educational and organizational layer. When a case exceeds the protocols, or a red flag appears ,PALM stops giving information and explicitly directs the family to their doctor or emergency services.

How we built it

The backend was developed in Python to manage integrations between the different services and components of the platform. A dedicated team member worked on extracting and structuring relevant information from medical documents. The frontend was built using React and Next.js, providing a modern, responsive, and intuitive user experience.

Challenges we ran into

We encountered a problem in uploading embeddings in the weaviateDB, due to problem with a local embeddings model. The main challenge was parsing the medical document into structured data with OpenAI

Accomplishments that we're proud of:

We built an AI you can actually trust with your child: a closed-perimeter assistant constrained to validated pediatric guidelines, with every output traceable to its source and a hard-stop that routes families to a doctor when a case exceeds the protocols.
We turned the messy reality of family records- crumpled paper reports, photos, scattered PDFs- into one structured, searchable child health record.
We built both sides of the relationship in one product: an app that empowers the parent and a dashboard that gives the pediatrician a ready-made clinical picture.
We grounded the product in real clinical authority, with a neonatologist on the founding team and the problem validated directly against the NICU-to-home experience.

What we learned

Start where the pain and the data are densest. Pediatrics, and especially the NICU-to-home transition, is the sharpest version of the problem, which is why we begin there.
Trust is an architecture, not a disclaimer. Constraining the AI to validated protocols and engineering an explicit hard-stop is a harder product than an open chatbot, and it's the entire point.
The data already exists. Our job isn't to make families generate more. It's to stop them from losing what they already have.

What's next for PALM

Product. Move from hackathon prototype to a real MVP with families and pediatricians, starting from neonatal/NICU-to-home journey where the need is most acute. Expand the validated knowledge base and deepen the ingestion layer that turns messy real-world records into structured data.

How PALM makes money. Three reinforcing lines:

B2B: pediatrician membership. Pediatricians pay a subscription for the dashboard that lets them manage their patients better: pre-assembled clinical pictures, fewer repetitive out-of-visit questions, and a longitudinal view they can't get anywhere else. They're not just users, they're distributors. Each pediatrician brings the families they already care for.
B2C: Palm Premium for families. A free tier builds the record and demonstrates value; a premium tier unlocks the full assistant and advanced features.
B2B2C: Partner-sponsored access. Insurers, corporate-welfare platforms, and baby/parenting brands have a direct interest in offering PALM premium to their members and customers. We license PALM as a service they sponsor for their users- a healthier, better-managed pediatric population is valuable to all of them, and it puts PALM in families' hands at scale without per-user acquisition cost.

Go-to-market. Grow inside-out: pediatricians who see the value distribute PALM to their patient base -low-cost, high-trust acquisition, while partnerships open a parallel, scaled distribution channel. The long-term goal: PALM as the operating system for a child's health from day one, the foundation for a lifetime of better-managed care.