Описание вакансии
Responsibilities:
- Design, build, and operate production agentic workflows and the platform harnesses they run on (orchestration, tool integrations, shared context, extension points for other teams)
- Practice intent engineering: turn goals into clear specifications, rules, constraints, and acceptance criteria that AI systems can execute reliably
- Stay current on AI tooling and practices, and evangelize what works across Engineering so partners adopt proven patterns
- Engage hard in design discussions and healthy debate while the direction is open, then pivot with equal energy into implementation once the team decides
- Build observability and guardrails: tracing, regression detection, human-in-the-loop controls, safe rollout, and operability for agentic systems others depend on
- Own critical pieces of agent runtime: job isolation, scheduling, execution environments, secrets and access, and production operations in the cloud
- Design and operate event-driven architectures where it fits (queues, streams, webhooks, async job fan-out) so agentic systems stay scalable and loosely coupled
- Ship safely and often: keep agentic systems on automated CI/CD paths with progressive delivery and clear rollback
- Bring a builder mentality: dogfood what we ship
- Use the platform to unblock yourself and the team, surface sharp edges, and turn real usage into product and platform improvements
- Compress software delivery loops by improving how AI helps engineers write, test, review, debug, and validate changes in real repositories and pipelines
- Partner across Engineering to understand workflows, ship tools that fit how people work, and drive adoption through playbooks, examples, demos, and enablement
- Evaluate build vs. buy; stay current on models, agent frameworks, MCP-style tool protocols, and the broader AI tooling ecosystem
- Define and track success metrics for adoption, quality, reliability, and satisfaction
- Use feedback and data to decide where to invest
- Partner on security and data access so tools are useful while respecting permissions and company policy
- Take designed projects from zero to production: own the path from agreed design through implementation, rollout, and day-two operability
- This position may include stand-by, on-call, or off-hours duties for systems you own
Requirements:
- Staff-level (or equivalent) professional software engineering experience building and operating production distributed systems
- Strong TypeScript (and modern JavaScript) plus Node.js experience shipping production services
- Polyglot comfort is welcome, TypeScript is the primary stack for this team
- Strong software engineering fundamentals and the ability to ship quickly: design, testing, debugging, APIs/services, and code quality
- Experience with continuous deployment culture: automated pipelines, progressive delivery or safe rollout, monitoring and rollback, and a bias toward frequent, reversible production changes
- Hands-on fluency with modern LLM and agentic coding workflows in production or serious internal platforms
- Experience building backend services, integrations, automation, and internal tools
- Professional experience with agent orchestration, tool-calling systems, evaluation or guardrail techniques, and connecting agents to reliable backend systems
- Familiarity with agent frameworks, orchestration layers, and integrating external tools and data sources into LLM-based systems (MCP or equivalent experience is a plus)
- Experience with event-driven systems: queues, streams, pub/sub, webhooks, or similar async patterns in production
- Observability fluency: metrics, logs, traces, and using them to operate and improve production systems
- Clear communication, documentation, and teaching ability; comfort driving adoption, not only writing code
- Good judgment around security, permissions, data access, and safe tool rollout
- Ability to problem-solve from first principles, make sound trade-offs, and drive work independently through ambiguity
Nice to have:
- Hands-on Kubernetes in production
- Terraform or similar infrastructure-as-code for cloud provisioning
- Temporal or other workflow-platform experience
- Deeper AWS / cloud-native ops fluency (IAM, networking, running production workloads end to end)
Зарплаты и спрос: Program менеджмент
- Сейчас в каталоге 318 вакансий по этому направлению, удалённо можно работать в 20%.
- Середина вилки в этой вакансии, 223 000 $ в год, на 14% выше медианы по направлению (195 000 $). У половины вакансий от 172 000 $ до 218 000 $ (вакансий с вилкой: 37).
Все вакансии направления «Program менеджмент»
Контакты работодателя доступны по кнопке «Откликнуться» после входа.