We're looking for a skilled Staff AI Engineer with expertise in LLM/GPT-based agentic engineering and AI-native product development. The ideal candidate excels in a fast-paced, outcome-focused environment, designing AI-powered applications that leverage advanced reasoning, autonomous decision-making, and multi-agent collaboration.
Your tasks
Design, develop, and deploy AI-native applications leveraging Large Language Models (LLMs) and GPT-based architectures.
Implement multi-agent systems using frameworks like LangChain, AutoGen, CrewAI, and OpenDevin for autonomous reasoning and AI workflows.
Build and optimize retrieval-augmented generation (RAG) pipelines, utilizing vector databases (e.g., Pinecone, Weaviate, ChromaDB) for efficient knowledge retrieval.
Develop scalable LLM inference and fine-tuning pipelines using tools like vLLM, DeepSpeed, LoRA, and FSDP for cost-efficient model deployment.
Lead AI workflow automation, integrating LLMs with external tools, APIs, and real-world business logic.
Design and implement AI memory systems, combining short-term prompt memory with long-term persistent storage.
Establish best practices for AI observability, safety, and evaluation, implementing guardrails and continuous monitoring for AI reliability.
Optimize AI model serving through cloud-native deployments (Kubernetes, Ray Serve, BentoML), ensuring low latency and high throughput.
Drive CI/CD and MLOps best practices for AI models and prompts, ensuring reproducibility and continuous evaluation.
Your Profile
Experience building production-grade AI systems with LLMs (e.g., GPT-4, Claude, LLaMA, Mistral).
Strong background in software engineering with hands-on experience in Python, FastAPI, and LangChain.
Expertise in retrieval-augmented generation (RAG), integrating LLMs with vector databases (Pinecone, FAISS, Weaviate).
Familiarity with AI agent architectures and building multi-agent workflows or AI-powered automation systems.
Experience deploying and optimizing models using cloud-based deployments (AWS/GCP/Azure, Kubernetes, Ray, vLLM).
Knowledge of AI safety, evaluation, and observability, including guardrails (e.g., Guardrails AI) and LLM evaluation pipelines.
Ability to work in a fast-paced, outcome-driven environment, balancing experimentation with execution.
We offer
Be you at IU: We support and empower you to achieve your personal and professional goals. After all, your happiness significantly contributes to a positive and productive environment.
Work wherever you want...: Digitalization is part of our DNA. Decide for yourself which place sparks the most productivity in you. You love remote work? No problem! Would you like to work abroad for a while? Our WorkFlex Benefit makes it possible.
... and flexible in terms of time: It's your job; We trust you and give you the greatest possible freedom to organize yourself.
Take your knowledge to a new level: Where else, if not with us? Enjoy free access to all our e-learning platforms. In addition, we will finance you a complete technology course of your choice.
Rest well: We make an effort every day, but we also take the time to recharge our batteries: with 30 days of annual vacation plus an extra day of vacation on your birthday, as well as the 24th and the 31st of December.
Sustainably to your destination: The 9-euro ticket is back! We subsidize your DeutschlandTicket, so you only pay €8.70 per month. Do you prefer to travel by bike? Then a JobRad might be something for you. This way, you are mobile both professionally and privately and do something good for the environment.
Actions speak louder than words: We cherish our "Culture of Everyone" and would like to take this opportunity to empathize that we do not tolerate discrimination towards any race, gender, religion, age, sexual orientation or disability. We welcome you as you are and look forward to nurturing our diverse community at the IU!
Sounds like the perfect job for you? Then apply now! Simple, fast and even without a cover letter. If you have any further questions, please feel free to contact Anita at Anita.Streitferdt@iu.org