Location: Russian Federation
Format: remote
Duration: until the end of November with possible extension
Salary Range: from 1500 to 2000 per hour (rubles) without VAT - hourly payment
Work under an agreement with an individual entrepreneur (you must have IE status).
Architecture and Infrastructure Stack:
Kubernetes, Docker, Helm, Nginx/Ingress, Kafka/RabbitMQ, Redis, PostgreSQL, Object Storage, API Gateway, Vault, Keycloak/ADFS, Prometheus, Grafana, OpenSearch.
ML / MLOps Stack:
Python, FastAPI, vLLM, Triton, PyTorch, MLflow, ClearML, Airflow, GitLab CI/CD.
Integration Stack:
REST, WebSocket, OpenAPI, AsyncAPI, ETL, corporate IAM/SSO, LDAP/AD.
Requirements:
- Highload Services Architecture:
Experience in designing high-load distributed systems: horizontal scaling, fault tolerance, load balancing, caching, queues, rate limiting, graceful degradation.
- ML/LLM Services Architecture:
Experience in designing LLM service architecture for industrial use: inference-serving, batch/online processing, fvector storage, pipeline orchestration, GPU workload management.
- Integration into Corporate IT Landscape:
Experience in embedding ML services into corporate architecture: integration with API Gateway, IAM/SSO, corporate buses, monitoring and logging systems.
- Microservices and Event-Driven Architecture:
Understanding of microservices approach, SOA, event-driven architecture, async interaction, message brokers, and data stream processing.
- Containerization and Orchestration:
Proficient knowledge of Docker, Kubernetes, Helm, ingress/service mesh, principles of deploying and operating containerized solutions.
- Cloud and On-Prem Architecture:
Experience in designing solutions both on-prem and in private/public cloud, understanding of hybrid architecture, multi-zone deployment, disaster recovery, backup/restore, SLA/SLO/SLI.
- Integration Technologies and APIs:
Experience in designing REST//WebSocket integrations, interaction contracts, authorization and routing schemes.
- Security and Corporate Standards:
Knowledge of information security principles: TLS/mTLS, JWT, OAuth2, OIDC, SAML/ADFS, RBAC/ABAC, secrets management, requirements for protecting personal and sensitive data.
- Data and Storage:
Understanding of data storage and processing architecture: PostgreSQL/Redis/Object Storage.
- Observability and Operations:
Experience in designing monitoring and service support: metrics, logs, traces, alerts, capacity planning, performance profiling, incident/problem management. MLOps / DevOps practices: Understanding CI/CD for ML services, model registry, experiment tracking, reproducibility.
- Technical Stack and Languages:
Understanding of Python, SQL, Bash, Git. Ability to read code and design architecture considering framework and runtime limitations.
- Architectural Documentation:
Experience in preparing architectural artifacts: C4/UML/, integration diagrams, NFR, ADR, security/data flow diagrams.
- Experience and Level:
Practical experience as a System/Solution/Enterprise Architect for 3+ years, total IT experience for 5+ years. Experience with LLM products or data platforms.
Additional
● Experience in designing GPU infrastructure and inference pipelines for LLM/embedding/reranking services.
● Understanding of vLLM, Triton Inference Server, ONNX Runtime, MLflow, ClearML specifics.
● Experience in selecting architecture for different load profiles: low latency / high throughput / batch / streaming.
● Skill in evaluating the total cost of ownership for architecture: CAPEX/OPEX, cost of GPU/CPU/Storage/Network.
● Experience in designing multi-tenant platforms and isolation pipelines.
● Knowledge of FinOps / Capacity planning practices is a plus.
● Experience in passing architectural committees, coordinating solutions with information security, infrastructure, and corporate architects.
● Skill in formalizing non-functional requirements: performance, availability, security, maintainability, scalability.
● Understanding of corporate integration specifics: legacy systems, restricted environments.
