We are seeking an experienced AI/LLM Data Engineer to build and maintain the data pipeline for our Generative AI platform. The ideal candidate will be well-versed in the latest Large Language Model (LLM) technologies and have a strong background in data engineering, with a focus on Retrieval-Augmented Generation (RAG) and knowledge-base techniques. This role sits in the AI COE within DX Tech & Digital. As a AI/LLM Data Engineer (you will report into the Director, AI Solutions & Development who oversees the AI COE. You will work on highly visible strategic projects, collaborating with cross-functional teams to define requirements and deliver high-quality AI solutions. The ideal candidate will have a passion for Generative AI and LLMs, with a proven track record of delivering innovative AI applications.
Responsibilities
- Design, implement, and maintain an end-to-end multi-stage data pipeline for LLMs, including Supervised Fine Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) data processes
- Identify, evaluate, and integrate diverse data sources and domains to support the Generative AI platform
- Develop and optimize data processing workflows for chunking, indexing, ingestion, and vectorization for both text and non-text data
- Benchmark and implement various vector stores, embedding techniques, and retrieval methods
- Create a flexible pipeline supporting multiple embedding algorithms, vector stores, and search types (e.g., vector search, hybrid search)
- Implement and maintain auto-tagging systems and data preparation processes for LLMs
- Develop tools for text and image data crawling, cleaning, and refinement
- Collaborate with cross-functional teams to ensure data quality and relevance for AI/ML models
- Work with data lake house architectures to optimize data storage and processing
- Integrate and optimize workflows using Snowflake and various vector store technologies
Requirements
- Master's degree in Computer Science, Data Science, or a related field
- 3-5 years of work experience in data engineering, preferably in AI/ML contexts
- Proficiency in Python, JSON, HTTP, and related tools
- Strong understanding of LLM architectures, training processes, and data requirements
- Experience with RAG systems, knowledge base construction, and vector databases
- Familiarity with embedding techniques, similarity search algorithms, and information retrieval concepts
- Hands-on experience with data cleaning, tagging, and annotation processes (both manual and automated)
- Knowledge of data crawling techniques and associated ethical considerations
- Strong problem-solving skills and ability to work in a fast-paced, innovative environment
- Familiarity with Snowflake and its integration in AI/ML pipelines
- Experience with various vector store technologies and their applications in AI
- Understanding of data lakehouse concepts and architectures
- Excellent communication, collaboration, and problem-solving skills.
- Ability to translate business needs into technical solutions.
- Passion for innovation and a commitment to ethical AI development.
- Experience building LLMs pipeline using framework like LangChain, LlamaIndex, Semantic Kernel, OpenAI functions.
- Familiar with different LLM parameters like temperate, top-k, and repeat penalty, and different LLM outcome evaluation data science metrics and methodologies.
Preferred Skills
- Experience with popular LLM/ RAG frameworks
- Familiarity with distributed computing platforms (e.g., Apache Spark, Dask)
- Knowledge of data versioning and experiment tracking tools
- Experience with cloud platforms (AWS, GCP, or Azure) for large-scale data processing
- Understanding of data privacy and security best practices
- Practical experience implementing data lakehouse solutions
- Proficiency in optimizing queries and data processes in Snowflake or Databricks
- Hands-on experience with different vector store technologies
Benefits
US employees benefit package.