- Years of experience required 8
- Strong Python skills: Expertise in Python for data exploration analysis and development using libraries like Pandas Matplotlib and Scikit Learn. Experience with Jupyter Notebook for interactive coding.
- NLP expertise: Deep understanding of natural language processing concepts and experience using Hugging Face pipelines for tasks like text classification generation and entity extraction.
- LLM experience: Handson experience with LLM frameworks like LLamaIndex or Langchain to build semantic search retrievalaugmented generation (RAG) and hybrid search systems.
- Prompt Engineering: Ability to design and structure prompts for LLMs programmatically using APIs from OpenAI Vertex AI or Llama. Familiarity with common prompt engineering patterns.
- Vector Database knowledge: Experience with any vector databases like PineCone Qdrant Vespa or Weaviate for efficient similarity search and retrieval.
- NLP evaluation metrics: Familiarity with common metrics used to evaluate the performance of NLP models RAG systems particularly for retrieval and generation tasks.
- Azure AI Services experience: Experience with key Azure AI services such as Azure OpenAI Service Azure AI Search and Azure Document Intelligence for building intelligent solutions.
- Azure resource management: Proficiency in provisioning configuring and managing Azure AI resources including understanding cost management and security best practices.
- Azure DevOps: Experience integrating Azure AI services into continuous integration and continuous delivery (CI/CD) pipelines for efficient deployment and updates.
- Containerization: Experience with containerized deployments on Azure for scalable and portable AI solutions.
Nice to Have:
- Conversation AI platforms: Familiarity with conversation AI platforms like Kore AI RASA Google Dialogflow or CCAI for building conversational agents and chatbots.
- Approximate Nearest Neighbor libraries: Experience with libraries like FAISS or ANNOY for efficient approximate nearest neighbor search particularly for large datasets.
Advanced Prompting techniques: Understanding of advanced prompting techniques like ReAct Fewshot learning Chainofthought function calling and Responsible AI to enhance LLM performance and safety.