We're building the AI layer that healthcare workflows have always needed.
Curee is an AI workflow automation platform purpose-built for healthcare providers. We make it possible for health systems, physician groups, SNFs, and rehab facilities to automate the documentation, coordination, and care management tasks that consume clinical time — without replacing the humans who deliver care.
This is the Curee handbook. It's a living document that describes how we think, how we work, and what we're trying to build. It's not a marketing page — it's an honest account of what Curee is and who we are.
We publish this publicly for two reasons. First, because the people who choose to work with us — whether as customers, employees, or partners — deserve to understand what we actually stand for. Second, because we believe transparency produces better decisions, internally and externally.
If something here seems wrong, or you disagree with it, we genuinely want to hear from you.
The Curee team, San Francisco — 2025
Eliminate the administrative burden that separates clinicians from patients.
The average physician spends more than two hours on documentation and administrative work for every hour of direct patient care. For nurses, it's worse. For small physician groups and post-acute facilities operating on thin margins, the burden is existential.
Our mission is to make AI-powered automation so seamless that the clinical team barely notices it exists — except that they have more time, less stress, and better data.
We are not building tools that replace clinical judgment. We are building tools that clear away the noise around clinical judgment, so the humans doing this work can do what they're actually trained for.
“The documentation burden is not a minor inconvenience. It is a direct cause of burnout, diagnostic error, and worse patient outcomes. We think it's solvable.”
Healthcare's software stack is broken in a very specific way.
EHRs were built for billing, not care. They are extraordinarily good at capturing charges and generating records. They are extraordinarily bad at reducing the work required to do both.
The downstream consequences are well-documented: clinician burnout at record levels, prior authorization delays that cost lives, care coordination gaps that drive readmissions, and documentation that takes 6 clicks to accomplish what should take one.
Most AI tools in this space have tried to patch individual symptoms — a better ambient documentation tool here, a prior auth automation there. These are helpful. But they don't address the root issue: care teams don't need more point solutions. They need an intelligent layer that sits across their entire workflow and surfaces the right action at the right time.
2.1×
Hours of admin work per hour of patient care
54%
Of physicians report burnout as their primary challenge
$935B
Lost annually to administrative waste in US healthcare
The things we actually believe.
Not values as wall art. Values as decision filters — the things we invoke when we're choosing between two reasonable options.
Default to clarity
Healthcare is already complex. We refuse to add more noise. Every workflow, every interface, every output should reduce cognitive load — not add to it.
Work done right > work done fast
In clinical settings, a rushed output can cause real harm. We build systems that enforce accuracy by design, not by reminder.
Trust is earned, not assumed
We don't ask providers to take our word for it. We show our reasoning, surface our sources, and make it easy to override or question anything Curee produces.
The care team is the user
We design for the nurse charting at 2am, the hospitalist managing 40 patients, the SNF administrator buried in prior auth. Not for the IT buyer.
Automation should be invisible
The best AI workflows are the ones you barely notice — because they just work. We aim for zero-friction adoption.
Build in public (internally)
We over-communicate internally. Decisions, trade-offs, mistakes — everything is written down, shared, and open for challenge.
Principles for how we operate day-to-day.
We're a remote-first team with dense async communication and minimal synchronous overhead. Here's how that actually works.
Writing over meetings
We write to think, to decide, and to align. Every significant discussion starts as a doc. Synchronous time is reserved for what asynchronous communication genuinely can't solve.
Small teams, big ownership
We keep squads to 3–4 people maximum. Each person owns a surface end-to-end — from spec to deployment to monitoring. No ticket queues, no handoff debt.
Clinicians in the room
Every product squad has a clinical advisor who holds veto power on anything touching patient workflows. We do monthly embedded observation sessions at partner sites.
Ship, learn, iterate
We run on two-week cycles with hard commitments. Nothing sits in staging for more than a sprint. If it's not ready to ship, we don't build it yet.
Radical transparency on metrics
Every team sees every team's OKRs, adoption rates, and error logs. Hiding problems is the only thing we're intolerant of.
Our SF office — we're remote-first but gather once a quarter.
The people building Curee.
We're a team of clinicians, engineers, and designers who've spent time inside health systems and deeply regret the state of the software they use. We're fixing it.
Karthik Rajakumar
Founder and CEO
Karthik Rajakumar is the founder and CEO of Curee AI. He has spent more than 15 years at the intersection of healthcare, machine learning, and cloud infrastructure. Before founding Curee, he led healthcare AI and machine learning at Amazon Web Services, first as a principal architect for healthcare AI/ML and then as a practice leader for analytics and industry go-to-market, working with health systems, payers, and life-sciences organizations to put AI into production. Earlier he advised healthcare and life-sciences clients at Deloitte and PwC. He holds a master's in computer engineering from Syracuse University and bachelor’s in computer science from Anna University (India). He started Curee to close a gap he saw repeatedly in the field: clinical teams have the data and the judgment, but the administrative layer between them and the patient keeps growing.
Dr. Juewon Khwarg
Chief Medical Officer
Dr. Juewon Khwarg is the chief medical officer of Curee AI and a board-certified physiatrist who continues to practice. He trained in physical medicine and rehabilitation at the VA–UCLA program and completed a fellowship in sports, spine, and regenerative medicine. He earned his MD at Case Western Reserve University, his undergraduate degree at Princeton University, and an MBA focused on the business of medicine at Indiana University's Kelley School of Business. Alongside Curee, he serves as a regional physician leader at Medrina and has held medical-director roles across post-acute and rehabilitation care. He leads Curee's clinical design so that every workflow the product touches reflects how care is actually delivered, not how software assumes it is.
Hemed Kaporo
Chief AI Scientist
Hemed Kaporo is Curee's chief AI scientist. He holds a PhD in computer science from Sabanci University, where his research centered on natural-language processing, and has spent over a decade building machine-learning and NLP systems, including a research fellowship in machine learning and NLP at Huawei. His expertise spans healthcare information technology, biomedical NLP, and medical coding, the exact disciplines Curee's product depends on. At Curee he owns the parts of the system that have to be right every time: PHI redaction, model evaluation, and the prompt and retrieval architecture that keeps the AI grounded in each customer's own clinical guidelines.
Azzam Asghar
Head of Engineering
Azzam Asghar is Curee's head of engineering. He owns the systems that let Curee run securely alongside any EHR: the desktop workstation, the multi-OS readers, and the orchestration layer beneath them. A full-stack engineer with more than a decade of experience, he previously founded and ran his own software company, shipping production systems across SaaS, cloud, and DevOps. He holds a computer-science degree from the National University of Computer and Emerging Sciences (FAST-NUCES). He builds healthcare infrastructure that earns clinical trust through reliability rather than promises.
Skylar Stanley
Applied AI Architect
Skylar Stanley is an applied AI architect at Curee, focused on turning each customer's clinical and operational rules into the grounding and evaluation layer behind the workstation, and on holding model output to a clinician-reviewed accuracy bar before deployment. An NLP specialist, Skylar builds and optimizes transformer-based language models, with hands-on depth in fine-tuning and inference efficiency (LoRA, long-context methods, and KV-cache optimization). Earlier, as an AI engineer at Accenture, Skylar delivered NLP and machine-learning systems across healthcare, finance, and retail. Skylar holds a master's and a bachelor's in computer science from the University of California, San Diego.
Brandon Cano
Applied AI Architect
Brandon Cano is an applied AI architect at Curee. The applied AI architects own what makes Curee accurate for one specific customer: translating a facility's clinical guidelines, payer rules, and coding logic into the knowledge and evaluation layer the AI reasons over, then measuring every output against a clinician-reviewed standard before it reaches production.
We're hiring across engineering, clinical, and design.
We look for people who have strong opinions about the problem we're solving and the patience to do the unglamorous work required to solve it well. Healthcare is hard. Building AI in healthcare is harder. We're not looking for people who want to make a quick dent.
If you've spent time in health systems and know exactly what's broken, we want to talk. If you've built infrastructure that has to be right — in finance, defense, healthcare, or elsewhere — we want to talk. If you're a designer who's genuinely frustrated by the UX of clinical software, we want to talk.
Staff Software Engineer, Backend
Engineering · Full-time · Remote
Senior ML Engineer, Clinical NLP
AI / ML · Full-time · Remote
Product Designer
Design · Full-time · Remote
Clinical Informatics Lead
Clinical · Full-time · SF Preferred
Director of Customer Success
GTM · Full-time · Remote
Don't see your role?
We hire for potential and perspective. Send us a note.