Senior AI Developer (Automators)

Luxoft; Schweiz; Hybrid; CHF 130.000 - 180.000; Vollzeit

Zusammenfassung

Luxoft is seeking a Senior AI Developer to design and build AI-powered internal tools for QA automation. The role combines AI engineering with hands-on test automation across Android TV platforms, using Java, Appium, and UiAutomator2 in a hybrid setup.

You will architect RAG pipelines, develop Lambdas in Python, and integrate AI capabilities into CI/CD workflows. Strong expertise in LLM orchestration and prompt engineering is required.

Qualifikationen

• AWS Bedrock hands-on experience with model access and Lambda integration.
• Experience designing and operating AI agents with agentic workflows and multi-tool orchestration.
• End-to-end RAG pipeline implementation: chunking, embedding, indexing, retrieval, generation.
• Proficient in prompt engineering (zero-shot, few-shot, JSON outputs).
• Knowledge of vector databases (OpenSearch, Pinecone, Faiss).
• Attention to guardrails and mitigation of hallucinations in LLMs.
• Understanding of when to use Fine-tuning vs RAG approaches.
• Experience with LangChain or similar orchestration frameworks.
• Strong embeddings expertise and semantic similarity.
• Proficiency in Python for Lambda and data processing.
• Java-based test automation for Android (Appium/UiAutomator2).
• Android/ADB device management and test execution.
• Experience with ReportPortal or similar test reporting tools.
• REST API concepts and practical usage.
• Jenkins/CI-CD pipeline integration and debugging.
• AWS ecosystem (S3, Lambda, API Gateway, IAM, OpenSearch Serverless).
• Docker-based test execution environments.

Aufgaben

• Design and implement AI-powered solutions for automated test failure triage and CBT pipelines.
• Build end-to-end RAG pipelines and document ingestion workflows.
• Develop Python Lambda functions and API Gateway endpoints for AI capabilities in CI/CD.
• Apply prompt engineering best practices and guardrails; evaluate LLM accuracy.
• Leverage Cursor IDE and MCP integrations for AI-assisted test generation.
• Write and maintain automated test suites in Java (Appium/UiAutomator2).
• Develop functional, regression, non-functional, and CBT test suites.
• Triage test failures in ReportPortal and integrate results with QMetry (QTM4J).
• Support CI/CD health via Nightly builds, RC, and release automation in Jenkins.
• Contribute to framework improvements and participate in Kanban ceremonies.
• Present AI demos to stakeholders and document architecture and pipelines in Confluence.

Kenntnisse

AWS Bedrock

AI agents

RAG pipeline

Prompt engineering

Vector databases

LLM guardrails

Fine-tuning vs. RAG

LLM orchestration

Embeddings

Python

Java

Appium/UiAutomator2

Android/ADB

ReportPortal

REST API

Jenkins/CI-CD

AWS

Docker

Tools

OpenSearch Serverless

Pinecone

Faiss

LangChain

Kotlin

Jobbeschreibung

Project description

We are building and maintaining one of the largest OTT platform test automation frameworks, serving millions of customers across streaming TV platforms. The team develops a Java/Appium-based automation framework for Android TV devices and is actively expanding it with AI-powered tooling.We are looking for a Senior AI Developer. This is a hybrid role combining the design and development of AI-powered internal tools with hands‑on test automation engineering skills. The ideal candidate is a software engineer who understands both QA automation and modern LLM/RAG systems — and can translate test engineering problems into practical AI solutions.

Responsibilities

• Design and implement AI-powered solutions focused on:o Automated test failure triage — LLM + RAG pipeline classifying ReportPortal failures (logs, stack traces, screenshots) into structured categories (PRODUCT_BUG, AUTOMATION_BUG, SYSTEM_ISSUE) using AWS Bedrock + Claudeo AI-based Change-Based Testing (CBT) — LLM-driven test case selection using semantic similarity between code changes and test coverageo AI test case generation from feature specs, Jira tickets, and Confluence documentation
• Build and maintain end-to-end RAG pipelines: document ingestion → chunking → embedding → OpenSearch Serverless vector store → retrieval → LLM response generation
• Develop AWS Lambda functions (Python 3.12) and API Gateway REST endpoints to integrate AI capabilities into CI/CD pipelines
• Apply prompt engineering best practices (system prompts, structured JSON output, guardrails) and drive continuous evaluation of LLM solution accuracy
• Use Cursor IDE with MCP integrations, agentic workflows, and context/rules files to accelerate test code generation and maintenance
• Write, maintain, and expand automated test suites in Java (Appium / UiAutomator2) for Android TV platforms
• Develop and maintain functional, regression, NFR, and CBT test suites
• Triage and resolve test failures in ReportPortal; integrate AI triage results with QMetry (QTM4J)
• Support CI/CD pipeline health — participate in Nightly Build, RC, and release automation runs via Jenkins
• Contribute to framework codebase improvements — bug fixes, refactoring, enhancements
• Participate in Kanban ceremonies and PI planning under the ART team
• Present AI solution demos to stakeholders and engineering leadership
• Document AI system architecture, RAG pipelines, and tools in Confluence

SKILLS

Must have

• AWS Bedrock — hands‑on: model access, Knowledge Bases, Lambda integration (primary AI platform)
• AI agents & Agentic tooling — practical knowledge of designing and operating AI agents, including agentic workflows, reusable skills, rules/guardrails, commands, and multi-tool/multi-agent orchestration
• RAG pipeline — end-to-end implementation: chunking, embedding, vector indexing, retrieval, generation
• Prompt engineering — zero-shot, few-shot, chain-of-thought, structured output (JSON mode), multi-turn
• Vector databases — working knowledge of OpenSearch, Pinecone, or Faiss; understands vector vs. graph DB difference
• LLM guardrails — input/output filtering, hallucination mitigation strategies
• Fine-tuning vs. RAG — ability to reason through which approach fits a given problem
• LLM orchestration — LangChain, LangGraph, or LlamaIndex
• Embeddings — understands semantic similarity; experience with Amazon Titan Embed or equivalent
• Python — for Lambda functions, AI pipeline scripting, and data processing
• Java — 3+ years of hands‑on test automation development
• Appium / UiAutomator2 — mobile/Android UI automation
• Android / ADB — device management, test execution
• ReportPortal or equivalent test reporting tool
• REST API — concepts and hands‑on usage
• Jenkins / CI-CD — pipeline debugging and integration
• AWS — S3, Lambda, API Gateway, IAM, OpenSearch Serverless
• Docker — containerized test execution environments

Nice to have

• Cursor IDE advanced features — .cursorrules, memory-bank context files, MCP server integration, and agentic triage workflows
• Android TV platforms — STB / embedded device testing experience (Fire TV, Roku, or similar)
• QMetry (QTM4J) — test management integrated with Jira
• Streamlit — for building internal AI dashboards
• DSPy — programmatic prompt optimization
• AWS SageMaker / MLflow — model evaluation and experiment tracking
• Kotlin — for tooling alongside Java