Claire Kim

Claire Kim

Ph.D. Candidate · LLM / NLP · Open to 2026 industry roles

About

I am Claire (Nayoung) Kim, a Computer Science Ph.D. candidate at the School of Computing and Augmented Intelligence (SCAI) at Arizona State University. My research advances trustworthiness and truthfulness in AI, with the goal of building fair, robust, and accurate NLP and LLM systems — spanning fairness, bias mitigation, robustness, and human-centered evaluation.

I have worked as a research assistant at the DHS Center for Accelerating Operational Efficiency (CAOE) and the Data Mining and Machine Learning (DMML) Lab, advised by Dr. Huan Liu and Dr. Mickey Mancenido. I previously earned my B.S. and M.E. at Korea University, where I was advised by Dr. Jaewoo Kang.

Currently seeking full-time Applied Scientist / ML Engineer / Research Scientist roles starting Summer 2026.

Download Resume

Experience

  • Applied Scientist Intern, Amazon — Bellevue, WA · Fall 2025
  • AI/ML Intern, AMD — Austin, TX · Fall 2024 & Summer 2025
  • Research Assistant, DHS-CAOE — Tempe, AZ · 2022 – 2025
  • Research Assistant, ASU DMML Lab — Tempe, AZ · 2021 – 2022
  • Research Assistant, Korea University DMIS Lab — Seoul, KR · 2017 – 2019

Skills

Languages
Python, SQL, JavaScript
ML & DL
PyTorch, TensorFlow, Hugging Face (Transformers, PEFT, Accelerate, Datasets), Pandas, NumPy
Fine-tuning
Supervised fine-tuning (SFT) · classification · forecasting · RLHF · parameter-efficient methods (LoRA, P-tuning)
LLM Systems
RAG, vector DBs (Weaviate, Chroma, Faiss), LLM-as-a-judge, synthetic QA generation, multi-agent LLM, hallucination mitigation, prompt engineering, inference-time scaling
Frameworks
LangChain, LlamaIndex, Node.js, Flask, Streamlit
AI-Assisted Dev
Claude Code, Cursor, GitHub Copilot, agentic coding workflows, MCP (Model Context Protocol)
Production & MLOps
End-to-end pipelines, model serving, Python packaging, inference optimization (throughput / latency), AWS (SageMaker, S3, Redshift, Glue, Lambda), GCP, Docker, WandB, MLflow
NLP & Models
Claude, GPT-4o, Llama-3, Mistral, DeepSeek, BERT · topic modeling, sentiment & stance detection, text summarization
Methods & Tooling
Bayesian inference & uncertainty quantification · Git · Linux / shell · Jupyter

News

  • Sep 2025 — Joined Amazon as an Applied Scientist Intern (Bellevue, WA).
  • May 2025 — Returned to AMD as an AI/ML Intern (Austin, TX).
  • Mar 2025 — PADTHAI-MM paper accepted to AI Magazine.
  • Sep 2024 — Robust Stance Detection presented at ASONAM 2024.
  • Aug 2024 — Joined AMD as an AI/ML Intern (Austin, TX).

Selected Projects

Reasoning-Level Fairness in LLMs — Under submission
A reasoning-aware framework that mitigates bias inside intermediate LLM reasoning steps, not just final answers. Introduced the Reasoning Bias Rate metric and combined process supervision, fairness-aware RL, and inference-time guided decoding to substantially reduce bias while preserving baseline accuracy.
Tech: Python · PyTorch · OpenR · inference-time scaling · fairness-aware RL

Fair Language Modeling via Parameter-Efficient Methods — 2024
Reduced social biases in BERT and LLaMA for toxicity and hate-speech detection using reinforcement learning combined with parameter-efficient fine-tuning.
Tech: Python · PyTorch · Hugging Face · LoRA · RL

MASTOPIA: Transparency in LLM-Assisted Intelligence Analysis — Under review (Human Factors)
A multi-agent RAG system operationalizing MAST tradecraft standards, evaluated through a large human-subject study. Showed that adding transparency features did not improve performance and can induce overreliance — motivating adaptive, on-demand transparency.
Tech: Python · GPT-4 · RAG · multi-agent LLM · vector DB · human-subject study

See the full project list.

Selected Publications

PADTHAI-MM: A Principled Approach for the Design of Trustworthy, Human-Centered AI Systems Using the MAST Methodology
Myke C. Cohen, Nayoung Kim, Yang Ba, et al.
AI Magazine, 2025.
[pdf] [code]

Robust Stance Detection: Understanding Public Perceptions in Social Media
Nayoung Kim, David Mosallanezhad, Lu Cheng, Michelle V. Mancenido, Huan Liu
ASONAM, 2024.
[pdf] [code]

Debiasing Word Embeddings with Nonlinear Geometry
Lu Cheng, Nayoung Kim, Huan Liu
COLING, 2022.
[pdf] [code]

See the full publication list.