Pay Range
$120,000 – $180,000 per year
Transform cutting‑edge research into reliable, user‑visible AI features in a mental health context. This role focuses on turning prompts into production‑ready end‑to‑end LLM features for voice and text therapy sessions, prioritizing safety and evaluation to ensure quality outcomes.
Responsibilities
- Ship end‑to‑end LLM features for voice + text therapy sessions : prompt / agent design, tool use / function calling, latency & turn‑taking handling, and production deployment.
- Build offline and online evaluation loops (unit tests, regression suites, shadow / production checks) that track reliability and user outcomes (e.g., anxiety‑score trends like GAD‑7), and use them to decide what ships.
- Implement and iterate safety systems : risky‑response detection, fallback / deferral strategies, human‑in‑the‑loop escalation, and post‑incident reviews; treat safety as a first‑class product feature.
- Own Python services and lightweight product surfaces (internal tools, small UX hooks) that speed up experimentation and founder feedback loops.
- Partner with Design (voice / chat UX) and Clinical Research to translate findings into product improvements and safeguards; instrument what you ship so we can learn quickly in production.
- Drive a weekly shipping cadence : prototype → evaluate → harden → release; document decisions and metrics so the team can build on them.
Qualifications
Strong Python plus hands‑on prompt / LLM engineering (tool use, function calling, retrieval or memory patterns, evals); you’ve shipped something real users touched.Product sense and speed : you can simplify ambiguous problems, choose pragmatic baselines, and deliver value in days—not quarters.Track record of safety‑critical thinking (red‑flag detection, guardrails, paths) and comfort being accountable for quality in production.Evidence‑driven mindset : you instrument features and are comfortable tying quality to measurable outcomes (not just engagement).Collaboration in a tiny, high‑trust team : you like building in person with founders and cross‑functional partners (design / research). In‑person, San Francisco required; U.S. work authorization (no sponsorship needed).Ideal Candidate
We’re looking for a design–engineering hybrid who owns discovery → UX / UI → implementation for our iOS app, iterates quickly with the founders in person in San Francisco, and cares deeply about measurable outcomes and safety.
What Makes You a Strong Fit
You’ve shipped mobile product end‑to‑end (portfolio shows discovery, design, and hands‑on build), ideally in Swift / SwiftUI.You can design stateful conversational experiences (voice + chat) and instrument what you ship to learn quickly.You use research and data to decide—e.g., you can explain how you’d evaluate changes with validated measures like GAD‑7, not just engagement metrics.You thrive on small‑team pace and high ownership, collaborating daily with founders in person in SF.You treat safety as a product feature and can describe guardrails you’ve designed for sensitive contexts.Design‑hungry, ideally with some engineering skills or interest; new‑grad OK.Seniority Level
Associate
Employment Type
Full‑time
Job Function
Information Technology
Industries
Hospitals and Health Care and Technology, Information and Media
#J-18808-Ljbffr