AI Expert – Generative AI Solutions Developer

About the role

We are seeking an experienced AI Expert who can independently design, develop, and deploy Generative AI solutions and embed AI features into Bosch products. The role involves end‑to‑end engineering of LLM/SLM‑based applications and agentic systems on Azure or AWS, from proof‑of‑concept through production deployment.

Key responsibilities

• Design, develop, and deploy production‑grade Generative AI solutions using Large and Small Language Models.
• Build and optimise Retrieval‑Augmented Generation pipelines, including chunking, embeddings and vector‑database integration (Azure AI Search, ChromaDB, FAISS, Pinecone).
• Develop multi‑agent and single‑agent systems with frameworks such as LangGraph, LangChain, AutoGen or CrewAI, including tool integration, memory and state management.
• Deploy solutions on Azure (Azure OpenAI, AI Studio, AI Search) and/or AWS (Bedrock, SageMaker), containerise with Docker and implement CI/CD pipelines.
• Apply MLOps/LLMOps practices for versioning, evaluation, observability, cost monitoring and guardrails.
• Collaborate with product, software, QA and architecture teams to embed AI capabilities into Bosch products while ensuring security, privacy and responsible AI compliance.

Required profile

• 10–15 years overall software/IT experience with 5–7 years hands‑on Generative AI experience.
• BE/BTech/MCA/MSc/MTech in Computer Science, IT, AI/ML or related field.
• Proven ability to work independently end‑to‑end and solve complex problems.
• Strong communication and stakeholder‑management skills.

Required skills

• Python programming and REST API development.
• Experience with Azure and/or AWS AI services (Azure OpenAI, Azure AI Search, AWS Bedrock, SageMaker).
• Agentic AI frameworks: LangGraph, LangChain, AutoGen, CrewAI.
• Retrieval‑Augmented Generation, prompt engineering, embeddings and vector databases (ChromaDB, FAISS, Pinecone).
• Docker, Git, CI/CD pipelines, containerisation.
• MLOps/LLMOps, model fine‑tuning, LiteLLM Proxy.