We are the Platform V Synapse Event Mesh team. We work on a high-load platform where reliability, security, and development speed are not just words, but daily practice.
We are currently looking for an Automation Engineer who will take ownership of the internal side of our processes: from processing vulnerability and bug tickets to controlling SLA and reporting. Not a manager who will "kick" teams, but an engineer who will replace routine with code and agents.
If you get annoyed doing the same thing manually more than twice, you should join us.
Responsibilities:
- Primary ticket analysis agent: classification, vulnerability triage, routing to the team, suggesting a responsible person
- Autonomous self-learning agent on a self-hosted framework (OpenClaw, Hermes, or similar) with a dashboard for visualizing and analyzing erroneous verdicts
- Regression loop for the agent: a labeled sample of tickets on which every change of prompt, model, or skill set is run
- Generation of FAQs and a knowledge base from the history of resolved tickets
- Transition of manual ticket processing (vulnerabilities, bugs, standards) into automated pipelines
- Control of ticket inclusion in release trains
- Automatic distribution of status dashboards via email and messengers
- Working on metrics and improvements (identifying bottlenecks in processes and eliminating them with measurable impact)
- R&D and experiments (predictive analytics: predicting SLA violations before they occur).
Requirements:
- Python: 2–3 years of commercial development. FastAPI, Pydantic, one of the agent orchestration frameworks (LangChain / LangGraph, LlamaIndex, Pydantic AI, or similar), structured output validation libraries
- Integrations between systems not only via REST: webhooks, gRPC, GraphQL, message queues (Kafka, RabbitMQ, NATS), JSON-RPC, scheduled and direct database dumps, legacy SOAP. Understanding when a webhook is needed, when polling, and when a queue is needed.
- Asynchronous operations and working with unreliable external APIs: timeouts, retries, partial failures
- Local models and understanding where they are more appropriate than external LLM APIs (OpenAI, Anthropic, YandexGPT, or similar)
- Local models and understanding where they are more appropriate than external LLM API (OpenAI, Anthropic, YandexGPT, or similar)
- DevOps: CI/CD, Git, Docker, Kubernetes
- Confident work with Linux (RHEL, CentOS, AltLinux)
- Willingness to say "LLM is not needed here, a script will suffice"
Conditions:
- Hybrid work format
- Annual bonus and salary review
- Accredited IT company status
- Extended voluntary medical insurance from day one and preferential family insurance
- Sberbank Corporate University, internal educational platform, participation in IT conferences
- 90 days of remote work from any region of the Russian Federation
- Preferential mortgage at Sberbank, Prime subscription with the ability to share with three close relatives.
