LLM Researcher

Role Overview:

• Build and deploy LLM-powered systems for dynamic, real-time sports commentary and interactive avatars.
• Work with cutting-edge frameworks, RAG, and multimodal models.

Key Responsibilities:

• Fine-tune and train LLMs, integrate tools, and optimize inference.
• Develop RAG systems and deploy them into production with monitoring.
• Ensure structured output generation and agentic reasoning workflows.

What We're Looking For:

• Solid experience with LLMs in production, including training and fine-tuning.
• Proficiency in Python, PyTorch, Hugging Face Transformers, and FastAPI.
• Experience with vector or graph databases and model serving tools like vLLM, Triton, or DeepSpeed.

Nice to Have:

• Sports broadcasting or data-driven content generation experience.
• Familiarity with multimodal LLMs and open-source contributions or research in NLP/agent systems.