AI Research Engineer

At LlamaIndex, the AI Research Engineer will join the core document understanding team to push the state-of-the-art in document AI. This role focuses on building accurate, cost-efficient APIs that differentiate us in a competitive market. You’ll collaborate across data, research, and product to translate ambitious research into reliable customer-facing capabilities.

What youll do

• Develop and train vision-language models for document processing, including data curation, synthetic data generation, and benchmark building.
• Evaluate base models and perform post-training to meet performance benchmarks.
• Iterate on the agent layer on top of base models to improve quality and cost-effectiveness.
• Collaborate with customers (less than 15% of the time) to translate product needs into benchmarks and model capabilities.
• Build on top of our existing foundation to improve experimentation velocity and develop novel methods to beat the status quo.

What LlamaIndex is looking for

• 3+ years of experience in AI research engineering roles.
• Experience with document AI or agentic systems, including vision models and document processing.
• Experience with benchmarking and training ML models.
• Experience with ML model serving and optimization tools (e.g., vLLM, TensorRT, ONNX).
• Experience with AI agent workflow tools (e.g., Claude Code, Cursor).
• Experience working in early-stage startup environments or academic research.
• Proficiency in Python and PyTorch.
• Experience with containerization and orchestration (e.g., Docker, Kubernetes).