abra is seeking for an MLOps Engineer . This role requires Senior ML Platform Engineer (MLOps) to maintain and improve the infrastructure that turns AI into reliable, secure, production-grade capabilities in an air-gapped environment. You'll maintain and develop the platform layer for modern AI\ML systems: training and fine-tuning pipelines, model serving, evaluation infrastructure, experiment tracking, model registries, vector databases as part of RAG workflows, dataset versioning, and GPU orchestration with a strong focus on open-weight LLMs and mission-critical inference where data, models, and infrastructure remain fully controlled by the organization. This is a hands-on role for someone who understands both ML/AI workflows and production infrastructure, and who can partner with AI developers, data scientists, security teams, and product engineers to upkeep a low-friction yet highly secure AI platform.
Requirements: 5+ years in software, platform, DevOps, infrastructure, data, or ML engineering, with hands-on experience building or operating a production ML/AI platform. Kubernetes (or OpenShift/Rancher), containers, Linux, networking, storage, and production operations. Scripting skills and comfort collaborating with ML engineers and software developers. Model serving: vLLM / Triton Inference Server / KServe / BentoML / TorchServe / Ray Serve or similar. ML lifecycle: MLflow / Kubeflow / Metaflow / Airflow / Argo Workflows or similar. RAG & retrieval: Postgres/pgvector, OpenSearch/Elasticsearch, MongoDB Vector Search, Milvus, Weaviate, Qdrant, or similar. CI/CD & IaC: Argo CD, Azure DevOps Pipelines, Tekton, GitLab CI, or Jenkins; Terraform, Ansible, Helm, Kustomize, or similar. Solid grasp of the model lifecycle: experiment tracking, model registry, dataset versioning, evaluation, deployment, monitoring, LLM observability, rollback, and lineage. LLM gateways & edge proxies: NGINX (TLS, streaming, rate limiting) in front of an AI gateway such as LiteLLM, Bifrost, Portkey OSS, Helicone, Kong AI Gateway, Apache APISIX, Envoy AI Gateway, custom FastAPI or similar. Strong cross-layer debugging skills across infrastructure, application, data, model, and GPU layers.
This position is open to all candidates.