Hi there, I'm You Lyu 👋

MS in Computer Science @ New York University

Interests: Machine Learning, Open Source Software, System Automation & Architecture

---

🛠 Skills

• Languages: Python, C++, Java
• Machine Learning: PyTorch, NumPy, SciPy
• Engineering & Architecture: API Design, System Architecture
• Testing & Workflows: pytest, Git/GitHub Workflows, CI/CD

---

🏛 Publications

EchoMind: An Interrelated Multi-level Benchmark for Evaluating Empathetic Speech Language Models

ICLR 2026 (Poster) | 3rd Author

Engineering & Data Contributions:

• Data Engineering: Synthesized and processed a substantial volume of high-quality EQ-related dialogue data utilizing an automated LLM TTS voice cloning pipeline, coupled with rigorous manual filtering.
• Evaluation Framework: Designed the multi-level EQ evaluation metrics and programmatically synthesized 100% of the Multiple Choice Questions (MCQs) for the benchmark.
• Infrastructure & Deployment: Participated in the deployment, device handling, and comprehensive evaluation testing of selected baseline PyTorch speech models.

---

💻 Projects

🗣 High-EQ Spoken Dialogue Model (Active)

Collaborator @ Human Language Technology Lab (HLT Lab), CUHKSZ

• Partnering with the HLT Lab to train a next-generation speech language model that utilizes both explicit conversational context and implicit acoustic vocal cues (e.g., emotion, sound events) for human-like empathetic interactions.

🤖 Autonomous LLM Evaluation Agent Workflow (Experimental)

Lead Developer

• Designed and built an agentic workflow to dynamically construct high-quality benchmarks, solving the latency between rapid LLM iterations and static benchmark updates.