Senior GenAI Engineer

Luxoft Poland; Polska; Sur place; PLN 180 000 - 300 000; Plein temps

Résumé du poste

Luxoft Poland seeks a hands-on GenAI Engineer to design and deliver production-grade AI-powered solutions for an insurance client. You will build scalable Python/FastAPI backends, integrate LLMs (OpenAI GPT, Anthropic Claude) and extend MCP servers to expose enterprise tools to AI agents.

You will design RAG pipelines, agentic workflows, and observability across the stack, ensuring security and data privacy while collaborating with stakeholders to turn requirements into working solutions.

Qualifications

• 4+ years of professional software development experience focusing on Python.
• Experience building production REST APIs with FastAPI or similar frameworks.
• Hands-on experience integrating LLMs (OpenAI GPT, Anthropic Claude) into real apps.
• Practical MCP server (FastMCP preferred) and LLM tool ecosystems (Skills, Plugins).
• Proven experience designing and implementing RAG pipelines.
• Experience building agentic workflows with LangChain, LangGraph, LlamaIndex or similar.
• Hands-on with at least one vector DB (e.g., Pinecone, Weaviate, Qdrant, Chroma, pgvector, Azure AI Search).
• Experience with observability/diagnostics for LLM systems (LangSmith, Langfuse, Arize Phoenix, OpenTelemetry).
• Experience deploying/running apps in production: Docker, CI/CD, and major cloud (AWS/Azure/GCP).
• Daily use of VS Code and AI coding assistants (Claude, Codex).
• Strong software engineering practices: clean code, testing, code reviews, Git.

Responsabilités

• Design, develop and maintain scalable backend services and APIs using Python and FastAPI.
• Integrate LLM-powered features with OpenAI GPT and Anthropic Claude in real apps.
• Develop MCP servers (FastMCP) to expose tools/data to AI agents.
• Create Skills and Plugins to extend LLM capabilities for workflows.
• Design and implement Retrieval-Augmented Generation (RAG) pipelines with ingestion, embedding and retrieval.
• Build agentic workflows using LangChain, LangGraph or similar frameworks.
• Work with vector databases to enable semantic search and knowledge retrieval.
• Implement observability and diagnostics for LLM apps: tracing, logging, evaluation, token costs, latency.
• Own full lifecycle: development, testing, deployment and production support.
• Use AI-assisted tools (Claude, Codex) to accelerate delivery while keeping code quality.
• Collaborate with stakeholders to translate requirements into working solutions.
• Ensure enterprise security, data privacy and responsible AI standards.

Connaissances

Python

FastAPI

OpenAI GPT

Anthropic Claude

MCP servers

FastMCP

RAG pipelines

LangChain

LangGraph

LlamaIndex

Vector databases

Pinecone

Weaviate

Qdrant

Chroma

pgvector

Azure AI Search

Docker

CI/CD

AWS

Azure

GCP

VS Code

Codex

Git

English (B2+)

Security & data privacy

Outils

Kubernetes

Terraform

Bicep

LangSmith

Langfuse

Arize Phoenix

OpenTelemetry

Description du poste

We are seeking a hands-on GenAI Engineer to join a team developing production-grade, AI-powered solutions for a major US insurance provider. The role involves designing and delivering end-to-end LLM-based applications, including scalable backend APIs, agentic workflows, deployment, monitoring, and observability. You will support our client in applying generative AI to key insurance processes such as underwriting, claims, and customer service.

Responsibilities

Design, develop and maintain scalable backend services and APIs using Python and FastAPI.

Build and integrate LLM-powered features using OpenAI GPT and Anthropic Claude models.

Develop and maintain MCP (Model Context Protocol) servers using FastMCP to expose enterprise tools and data to AI agents.

Create and extend Skills and Plugins that enhance LLM capabilities for business-specific workflows.

Design and implement Retrieval-Augmented Generation (RAG) pipelines, including document ingestion, chunking, embedding and retrieval strategies.

Build agentic solutions using LLM orchestration frameworks such as LangChain, LangGraph or similar.

Work with vector databases to support semantic search and knowledge retrieval.

Implement observability and diagnostics for LLM applications: tracing, logging, evaluation, token and cost tracking, latency and quality monitoring.

Own the full application lifecycle: development, testing, deployment and production support.

Use AI-assisted development tools (Claude, Codex) to accelerate delivery while maintaining code quality.

Collaborate with client stakeholders, architects and business analysts to translate requirements into working solutions.

Ensure solutions meet enterprise standards for security, data privacy and responsible AI use.

Must have

4+ years of professional software development experience with a strong focus on Python.

Solid experience building production REST APIs with FastAPI (or a comparable framework).

Hands-on experience integrating LLMs (OpenAI GPT, Anthropic Claude) into real applications, including prompt engineering, tool/function calling and structured outputs.

Practical experience with MCP servers (FastMCP preferred) and LLM tool ecosystems (Skills, Plugins).

Proven experience designing and implementing RAG pipelines.

Experience building agentic workflows with orchestration frameworks such as LangChain, LangGraph, LlamaIndex or similar.

Hands-on experience with at least one vector database (e.g., Pinecone, Weaviate, Qdrant, Chroma, pgvector, Azure AI Search).

Experience with observability and diagnostics for LLM systems (e.g., LangSmith, Langfuse, Arize Phoenix, OpenTelemetry).

Experience deploying and running applications in production: Docker, CI/CD, and at least one major cloud platform (AWS, Azure or GCP).

Daily, confident use of modern development tools including VS Code and AI coding assistants such as Claude and Codex.

Strong understanding of software engineering best practices: clean code, testing, code review, version control (Git).

Upper-Intermediate (B2) or higher English, with the ability to communicate directly with US-based stakeholders.

Nice to have

Experience in the insurance or broader financial services domain.

Knowledge of LLM evaluation techniques (automated evals, LLM-as-judge, guardrails).

Experience with Kubernetes and infrastructure-as-code (Terraform, Bicep).

Familiarity with data privacy and compliance requirements for handling PII in regulated industries.

Experience with asynchronous Python, message queues or event-driven architectures.

Frontend experience (React, Streamlit) for building internal AI tools and demos.