NoFluffJobs Stacjonarnie Mid

Software Engineer (Backend)

Bayer

⚲ Warszawa

14 000 - 21 000 PLN (PERMANENT)

Wymagania

  • Python

Opis stanowiska

O projekcie: You’ll build and extend the backend of myGenAssist, an enterprise AI assistant platform. The backend is a FastAPI application that orchestrates LLM interactions, manages agent workflows with LangChain/LangGraph, handles document processing pipelines, and exposes tools via MCP (Model Context Protocol). Wymagania: Qualifications & Competencies (education, skills, experience): - 3+ years with Python in production backend systems - Strong FastAPI experience (async/await, dependency injection, middleware) - SQLAlchemy or SQLModel with PostgreSQL - Redis for caching and task queues - Experience integrating LLMs via API (OpenAI SDK or similar) - Testing with pytest - Docker for development and deployment - Understanding of async Python (asyncio, concurrent session management) - Git and PR-based workflow - Nice to Have: - LangChain / LangGraph for agent orchestration - Vector databases (Qdrant, pgvector) and RAG patterns - MCP (Model Context Protocol) - Document processing pipelines (Docling, Unstructured) - OpenTelemetry and distributed tracing - Kubernetes basics (we run on EKS) - AWS services (S3, Secrets Manager) - OAuth2/JWT authentication - Agile/Scrum experience Codzienne zadania: - Design and implement async APIs using FastAPI, following our patterns for session management, API versioning, and domain-driven structure. - Build and evolve AI agent workflows using LangChain and LangGraph — tool integration, state persistence, multi-step reasoning. - Design and optimize database schemas with SQLModel/SQLAlchemy and PostgreSQL, including migrations and vector search (pgvector, Qdrant). - Build background job pipelines with RQ (Redis Queue) for document processing, knowledge base indexing, and long-running tasks. - Implement observability using OpenTelemetry, Langfuse, Prometheus, and Grafana. - Write tests using pytest with parallel execution, covering unit, integration, and RAG evaluation. - Enforce code quality through reviews and adherence to security best practices (OWASP).