MLOps (Bank)

We are creating an ecosystem of digital employees based on advanced open LLM/VLM models. Our AI agents automate key operational processes, directly impacting the efficiency and competitive advantage of our clients' businesses.

Company: Selecty Position: MLOps (Bank) Income Level: 250-340k gross Format: Employment according to the Labor Code of the Russian Federation Remote in Russia

What you will be doing:

You will become a key engineer who transforms research prototypes into a fault-tolerant GenOps factory of digital agents.

Your tasks: DevOps for AI: Design and develop self-service CI/CD pipelines for internal deployment of AI agents and multi-agent systems.

Inference Management: Deploying and optimizing LLMs/VLMs during development (memory optimization, speed formation) using vLLM, Triton Inference Server, SGLang.

Reliability: Ensuring high service availability, working with horizontal pod autoscaling, and managing agent state.

360° Monitoring: Implementing comprehensive observation: from standard metrics (CPU/GPU, latency) to specific ML metrics (generation quality, data drift, response sentiment analysis).

Mandatory Requirements

Experience deploying LLM solutions to production.

• Proficient in Python and asynchronous programming (asyncio).

• Proficient in Kubernetes (Helm, cluster management).

• Knowledge of Gitlab CI, Jenkins, etc.

• Understanding of RAG system architecture, embedding model operation, and ranking.

• Ability to optimize inference: quantization, continuous batching, PagedAttention.

• Bonus points: experience with graph databases and building multi-agent systems

  Tech Stack: Python, FastAPI, langchain/llamaindex/haystack k8s, docker, S3, postgres Experience with vector and graph databases.

Conditions

Employment according to the Labor Code of the Russian Federation in an IT-accredited company DMS (including dentistry) Fitness compensation Discounts on English language learning (Skyeng) Salary indexation Modern equipment Company IT accreditation