DevOps Engineer

Managing and developing our Kubernetes platform across multiple clusters and environments including production, development, on-premises and public cloud.

Designing and overseeing hybrid cloud infrastructure across on-premises and public clouds (such as GCP, AWS), including workload placement, cross-cloud networking, and unified resource management.

Taking responsibility for the end-to-end CI/CD and GitOps process, including container build pipelines, image optimization, and progressive delivery using tools like ArgoCD/FluxCD.

Taking charge of the observability stack to provide a comprehensive view across all clusters using tools like Grafana, Mimir, Tempo, Loki, Pyroscope, OnCall, Prometheus, and supporting agent-assisted SRE workflows.

Managing and enhancing our inference platform, including vLLM serving and AIBrix for multi-model orchestration and autoscaling with a fleet of NVIDIA GPUs.

Operating platform services such as Kafka, Redis, PostgreSQL, OpenSearch.

Managing identity and access management with Keycloak integrated with Google Workspace, strengthening SSO, RBAC, and secrets management across the platform.

Strengthening network security across private load balancers, firewalls, and VPC segmentation and designing and maintaining hub-and-spoke/multi-AZ topologies.

Supporting training infrastructure with self-service VM provisioning, RunPod burst capacity, and Weights and Biases integration.

Driving infrastructure reliability, cost efficiency, and capacity planning as the platform scales.

Strong production experience with Kubernetes, including workloads and controllers, networking, storage, RBAC, and autoscaling. Familiarity with both cloud-managed and on-premises/self-managed Kubernetes is a plus.

Design-level networking experience, including the ability to defend tradeoffs in real network topologies such as hub-and-spoke, multi-AZ/multi-VPC, and equivalent enterprise patterns. Comfortable with VPCs, firewalls, load balancers, private cluster architecture, DNS, and routing. On-premises networking experience is a strong plus.

Proficiency in building and optimizing Dockerfiles and owning full CI/CD pipelines. Experience with CI/CD changes when deploying to Kubernetes is a bonus.

Previous experience in setting up and operating a full observability stack in production, including metrics, logs, traces, and alerting. Familiarity with the Grafana stack is a strong plus.

Comfort with SSO and identity, including integrating tools with a central IdP.

Strong Linux proficiency, infrastructure-as-code, and configuration management skills.

An ownership mindset and familiarity with operating at high-ownership environments.

Hands-on experience with Kafka, Redis, PostgreSQL, or OpenSearch at a production scale is optional but valuable.

Bonus points for experience with OpenStack and KVM virtualization, familiarity with vLLM internals, a background in AI/ML infrastructure or GPU cluster operations, experience with KEDA or event-driven autoscaling patterns, prior open-source contributions, and kernel-level Linux debugging and performance tuning.

Join a renowned research team to work on impactful projects

Take ownership of the core training code infrastructure used by the team

Engage with real models, data, and scale, rather than small-scale problems

Contribute to bridging the gap between research velocity and engineering quality

Enjoy a flexible work environment with a culture that values depth, clarity, and curiosity

PEGASI – IT Recruitment Consultancy | Email: recruit@pegasi.com.vn | Tel: +84 28 3622 8666