Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
ЗП не указанаразвернуть
На похожих ролях по рынку
$160к$198кмедиана$240к
50 в выборке
На СНГ рынке: $4.3к/мес ($3.6к - $5к)
92
Крутая вакансия
развернуть
Здесь можно развивать работающего в продакшене ИИ-агента с понятной зоной ответственности, строгой оценкой качества и сложными задачами голосового взаимодействия и страховых процессов; задачи разнообразны, но связаны с одним продуктом.
?Широкая зона задач
ИИ в продакшене
Современный ИИ-стек
Зрелая продуктовая компания
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
AI Engineer (Applied AI) (LLM/Voice AI/Insurance): Building and hardening a production omnichannel AI agent for first notice of loss with an accent on agent orchestration, retrieval, claims-system integrations, voice latency, and evaluation. Focus on designing reliable eval loops, analyzing real call traces, improving containment without sacrificing accuracy, and controlling synchronous voice interactions.
Location: United States — Remote
Company
hirify.global provides a cloud platform combining digital, core, analytics, and AI capabilities for property and casualty insurers.
What you will do
Build and own end-to-end slices of Concierge, including prompt and context architecture, retrieval over policy and coverage data, tool and API calls into claims systems, and human handoff.
Build and operate evaluation loops with golden sets, trajectory evaluation, LLM-as-judge, production call-trace analysis, and release gating.
Improve containment and resolution rates while preserving accuracy, and instrument the system to measure both.
Develop the real-time voice path, including turn-taking, interruption handling, latency budgets, LiveKit, and SIP telephony.
Use carrier calls and recordings to identify conversation failures and guide product improvements.
Requirements
Experience building and deploying AI applications, including an LLM application used by real users.
Fluency with LLMs, context engineering, RAG, and agentic workflows, supported by practical machine learning and deep learning expertise.
Ability to make unpredictable AI systems reliable through disciplined evaluations, error analysis, and statistical measurement.
Strong software engineering fundamentals, including reasoning about cost, latency, reliability, scalability, data stores, distributed systems, testing, security, privacy, and CI/CD automation.
Experience working with claims data, recorded calls, and personally identifiable information under carrier and regulatory scrutiny.
Daily experience using coding agents, including managing context, writing specifications when useful, creating verifiers, and deciding when to intervene.
Product sense and enough insurance context to shape agent behavior, prioritize MVP delivery, and recognize when claims workflows require additional rigor.
Ability and willingness to keep learning as AI tooling and practices evolve.
Nice to have
Experience with real-time or streaming voice systems, including ASR, TTS, telephony, SBC, or SIPREC.
Experience in insurance, claims, or another regulated domain where errors have material costs.
Experience building internal evaluation or observability tooling adopted by other engineers.
Experience with machine learning principles, model experimentation, and training and deploying models in production.
Experience with P&C insurance or hirify.global products and services.
Experience applying DSPy or GEPA prompt and pipeline optimization frameworks to production systems.
Culture & Benefits
Remote work in the United States.
Full-time employment with health, dental, and vision insurance, paid time off, and a company-sponsored retirement plan.
Some roles may include annual bonuses, commissions, or long-term incentive awards subject to applicable factors.
Equal opportunity and disability-inclusive employment with accommodations available during the hiring process.
Hiring process
Submit a short note describing something shipped where model unpredictability was the central problem and explaining how it was addressed.
Applications are subject to applicable criminal history and other background checks.
Будьте осторожны
Если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого. Это мошенники. Обязательно жмите «Пожаловаться» или пишите в поддержку. Подробнее в гайде →