Valarian
Valarian
„Valarian“ kuria duomenų suverenumo ir saugumo infrastruktūrą komercinėms įmonėms bei vyriausybinėms institucijoms. Jos platforma skirta gynybos, sveikatos priežiūros ir kitų reguliuojamų sektorių organizacijoms, kurioms reikia apsaugoti jautrią informaciją ir kartu laikytis vietos atitikties reikalavimų. „Valarian“ daugiausia dėmesio skiria duomenų lokalizavimui, detaliam prieigos valdymui, audituojamumui ir atsparumui kibernetinėms grėsmėms, tinklo sutrikimams, duomenų pažeidimams bei geopolitinei rizikai. Įmonė padeda tarpvalstybiniu mastu veikiančioms organizacijoms išlaikyti savo duomenų kontrolę nevaržant jų gebėjimo diegti naujoves ir konkuruoti tarptautiniu mastu.

Senior MLOps Engineer at Valarian

Run production LLM applications and agentic AI infrastructure at Valarian, a dual-use security technology company. Oversee Kubernetes deployments, observability, data pipelines, and service reliability.

Aprašymas

  • Own deployment, monitoring, orchestration, and reliability for production LLM applications and an agentic AI platform.
  • Design and operate LLM services, agentic applications, and supporting AI workloads on high-throughput Kubernetes clusters.
  • Automate container deployments, scalable inference endpoints, and blue-green or canary releases through CI/CD pipelines.
  • Maintain production agentic AI workflows, including multi-step tool execution, state management, and coordination between agents.
  • Investigate and resolve problems spanning AI application logic, orchestration frameworks, external APIs, and cloud infrastructure.
  • Develop monitoring and evaluation systems for output accuracy, drift, latency, cost, and hallucination metrics.
  • Add trace-level visibility into agent execution paths, including prompts, completions, tool calls, and individual steps.
  • Monitor and improve real-time and batch data pipelines, feature stores, vector indexes, and embedding updates.
  • Manage GPU and CPU resources, memory, autoscaling, and cloud costs to improve time to first token and throughput.

Reikalavimai

  • At least three years of experience in MLOps, LLMOps, platform engineering, AI infrastructure, or SRE focused on machine learning workloads.
  • Practical experience deploying and operating complex containerized applications at scale on Kubernetes, such as EKS, GKE, AKS, or bare metal.
  • Strong Python skills for backend automation, system debugging, custom tools, and integrations.
  • Experience with monitoring and logging tools such as Prometheus, Grafana, Datadog, or OpenTelemetry, plus LLM tracing and evaluation platforms such as LangSmith, Phoenix/Arize, Langfuse, or TruLens.
  • Familiarity with modern CI/CD systems such as GitHub Actions, GitLab CI, or ArgoCD, and infrastructure as code tools such as Terraform or Pulumi.
  • Strong operational understanding of agentic architectures, prompt chaining, function calling and tool execution, and non-deterministic application behavior.
  • Practical experience running Retrieval-Augmented Generation (RAG) pipelines and production vector databases is a plus.
  • Hands-on familiarity with agent development and orchestration frameworks is a plus.
  • Exposure to high-performance LLM serving engines and frameworks is a plus.
  • Understanding of AI security practices and rate limits or cost controls is a plus.

Privalumai

  • Equity.
  • Competitive salary.
  • Employer pension contributions.
  • Private health insurance.
  • Hybrid work arrangement.
  • Gym memberships.
  • Company retreats and meetups designed to support work-life balance.

Susiję darbai

Faculty

Senior Machine Learning Engineer at Faculty (Hybrid, UK)

Faculty

Lead the architecture and delivery of production-ready, responsible machine learning systems for Faculty’s clean energy and sustainability clients. Advise clients, guide deployment, and mentor engineers.

Atidaryti