Senior Principal AI Engineer, Agentic Systems at Charliehealth

Why Charlie Health? Millions of people across the country are navigating mental health conditions, substance use disorders, and eating disorders, but too often, they’re met with barriers to care. From limited local options and long wait times to treatment that lacks personalization, behavioral healthcare can leave people feeling unseen and unsupported. Charlie Health exists to change that. Our mission is to connect the world to life-saving behavioral health treatment. We deliver personalized, virtual care rooted in connection—between clients and clinicians, care teams, loved ones, and the communities that support them. By focusing on people with complex needs, we’re expanding access to meaningful care and driving better outcomes from the comfort of home. As a rapidly growing organization, we're reaching more communities every day and building a team that’s redefining what behavioral health treatment can look like. If you're ready to use your skills to drive lasting change and help more people access the care they deserve, we’d love to meet you. About the role The next generation of behavioral health will be built on agentic AI — systems that don't just predict or summarize, but act: scheduling, coordinating care, drafting clinical documentation, surfacing risk, and orchestrating the hundreds of micro-decisions that today consume clinician and operator time. Charlie Health's ambition is to build the most advanced agentic systems in healthcare. We are investing in a multi-agent architecture grounded in a real-time clinical signal graph, governed by HIPAA-native infrastructure, and deployed against measurable patient outcomes in a national virtual IOP. We are hiring a Senior Principal AI Engineer to lead the design, development, and deployment of these systems across our platform. This is a hands-on technical leadership role with deep cross-functional reach into Engineering, Product, Data, Clinical, and executive leadership. Responsibilities • Serve as the technical owner of agentic AI at Charlie Health, from architecture and model strategy through production deployment and ongoing reliability. • Lead the development of clinical operations agents that handle scheduling, intake, insurance verification, referral triage, and care coordination, replacing brittle human handoffs with auditable, supervised AI workflows. • Lead the development of clinician copilots for real-time documentation, risk surfacing, and treatment-plan support, integrated with our EHR and clinical data platform. • Define and build the agentic substrate: eval frameworks, observability, guardrails, tool registries, prompt and policy versioning, PHI-safe logging, and simulation environments. • Own the model strategy across frontier models, fine-tunes, distilled specialists, and classic ML, and articulate it clearly to executive and clinical leadership. • Partner with Product, Clinical, and Data teams to identify high-leverage opportunities for AI across patient and operational experiences. • Establish best practices around AI reliability, safety, evaluation, and explainability in a regulated clinical environment. • Mentor engineers and data scientists across the organization and help foster a strong culture of technical excellence and responsible AI innovation. What you'll work with • An executive team that has committed to an AI-native operating model and is funding it. • An AI Guild of senior engineers, data scientists, and clinical operators piloting AI across the company today. • A modern data foundation including Snowflake, dbt, pgvector, and an emerging Intelligence Layer that exposes patient and operational signals as first-class events. • Live workstreams in clinical documentation, referral-source conversion modeling, scheduling intelligence, and clinician productivity. • A national virtual IOP serving patients across 41 states, with the regulated data and outcome signal needed to train, evaluate, and improve real systems. Why this role is hard Behavioral health is one of the most demanding applied-AI environments. Patients are often in acute distress. The regulatory surface spans HIPAA, state-by-state telehealth and licensure law, and rising FDA scrutiny of clinical decision support. AI systems in this environment must be observable, reversible, and defensible to clinicians, to regulators, and to internal review. We are looking for an engineer who finds those constraints clarifying rather than limiting. Requirements • 10+ years of software engineering experience, with several recent years focused on LLM- or ML-powered systems in production. • Hands-on experience designing and operating agentic systems, including tool use, planning, multi-agent orchestration, eval design, and failure-mode analysis. • Track record of owning a complex technical initiative end-to-end, including evaluation infrastructure, observability, on-call, and regulatory or compliance review. • Strong communication skills and comfort operating across registers, from executive strategy conversations to engineering code review. • Genuine interest in healthcare and in applying AI responsibly in a regulated, high-stakes domain. • BS, MS, or PhD in Computer Science, Machine Learning, Engineering, Mathematics, or a related technical field, or equivalent practical experience. • Able to work a hybrid schedule of 4 days/week in our NYC office and located within 75 minutes' commuting distance of the office Benefits Charlie Health is pleased to offer comprehensive benefits to all full-time, exempt employees. Read more about our benefits here. Additional Information The total target base compensation for this role will be between $350,000 and $600,000 per year at the commencement of employment. Please note, pay will be determined on an individualized basis and will be impacted by location, experience, expertise, internal pay equity, and other relevant business considerations. Further, cash compensation is only part of the total compensation package, which, depending on the position, may include stock options and other Charlie Health-sponsored benefits. #LI-HYBRID Our Values • Connection: Care deeply & inspire hope. • Congruence: Stay curious & heed the evidence. • Commitment: Act with urgency & don’t give up. Please do not call our public clinical admissions line in regard to this or any other job posting. Please be cautious of potential recruitment fraud. If you are interested in exploring opportunities at Charlie Health, please go directly to our Careers Page: [Upgrade to PRO to see link] Charlie Health will never ask you to pay a fee or download software as part of the interview process with our company. In addition, Charlie Health will not ask for your personal banking information until you have signed an offer of employment and completed onboarding paperwork that is provided by our People Operations team. All communications with Charlie Health Talent and People Operations professionals will only be sent from @charliehealth.com email addresses. Legitimate emails will never originate from gmail.com, yahoo.com, or other commercial email services. Recruiting agencies, please do not submit unsolicited referrals for this or any open role. We have a roster of agencies with whom we partner, and we will not pay any fee associated with unsolicited referrals. At Charlie Health, we value being an Equal Opportunity Employer. We strive to cultivate an environment where individuals can be their authentic selves. Being an Equal Opportunity Employer means every member of our team feels as though they are supported and belong. We value diverse perspectives to help us provide essential mental health and substance use disorder treatments to all young people. Charlie Health applicants are assessed solely on their qualifications for the role, without regard to disability or need for accommodation. By clicking "Submit application" below, you agree to Charlie Health's Privacy Policy and Terms of Service. By submitting your application, you agree to receive SMS messages from Charlie Health regarding your application. Message and data rates may apply. Message frequency varies. You can reply STOP to opt out at any time. For help, reply HELP.