Yumo Xu

AI Researcher

Bellevue, Washington, United States4 yrs 10 mos experience

Key Highlights

Expert in post-training foundation models for AI.
Led development of LLM-powered enterprise solutions.
PhD in NLP with a focus on advanced reasoning.

Stackforce AI infers this person is a SaaS expert specializing in AI-driven solutions for enterprises.

Contact

Skills

Core Skills

Natural Language ProcessingMachine Learning

Other Skills

Post-training foundation modelsAsset generationPost-trainingMulti-turn RLSource attributionStreaming-based parsingEfficient snippet extractionAPI response compressionMulti-agent orchestrationSearch planningContext managementPythonResearchStrategic PlanningProject Management

About

I am an LLM researcher at Netflix. My overarching research goal is to build helpful and responsible AI systems to improve how people access information and take actions. My research centers on post-training foundation models to instill advanced reasoning and improve alignment. I am also interested in extending these capabilities to multi-modal domains and creating evaluations that meaningfully reflect their real-world utility. Prior to Netflix, I was a scientist at AWS AI Labs. I received my PhD in NLP at the University of Edinburgh, advised by Prof. Mirella Lapata.

Experience

4 yrs 10 mos

Total Experience

1 yr 5 mos

Average Tenure

5 mos

Current Experience

Netflix

Senior Research Scientist

Nov 2025 – Present · 5 mos · Remote

Post-training foundation models for asset generation

Post-training foundation modelsAsset generationNatural Language ProcessingMachine Learning

Amazon web services (aws)

2 roles

Senior Applied Scientist, AWS AI Labs

Sep 2025 – Nov 2025 · 2 mos · Bellevue, Washington, United States

• Led post-training for agentic search (multi-turn RL)

Post-trainingMulti-turn RLNatural Language ProcessingMachine Learning

Applied Scientist, AWS AI Labs

Apr 2023 – Sep 2025 · 2 yrs 5 mos · Bellevue, Washington, United States

Founding member of Amazon Q for Business / Amazon Quick Suite, an LLM-powered assistant for enterprises. Led the research and production of:
Source attribution: principle-driven citation eval (CiteEval, ACL'25), streaming-based parsing, efficient snippet extraction
Orchestration/Actions: API response compression for function calling (US Patent), built-in plugins, multi-agent orchestration
Agentic RAG: search planning, parallel/sequential tool execution, context management (featured in AWS Blogs: https://aws.amazon.com/blogs/machine-learning/bringing-agentic-retrieval-augmented-generation-to-amazon-q-business)

Source attributionStreaming-based parsingEfficient snippet extractionAPI response compressionMulti-agent orchestrationSearch planning+3