CV

Download CV (PDF)

Education

• M.S. in Applied Statistics & Data Science, University of California, Los Angeles (UCLA), Sep 2024 – present
  • GPA: 4.0 / 4.0
• B.S. in Data Science, University of California, San Diego (UCSD), Sep 2017 – Mar 2024
  • GPA: 3.56 / 4.0 (Upper Division: 3.85 / 4.0)

Work experience

• Samsung Electronics, AI Model Research Intern (Jun 2025 – Aug 2025)
  • Researched multimodal LLM merging to raise language-benchmark scores while keeping multimodal performance above target.
  • Designed layer-wise merge strategies and built a scalable safetensors merging pipeline supporting multiple models.
  • Evaluated merged models with LLMeval and LMMeval, improving language results without multimodal regression.
• Kirr LLC, Founder & Chief Growth Officer (Apr 2024 – present)
  • Founded a lifestyle brand, built its Shopify store, and led product positioning and early-stage growth strategy.
• Suntek Systems Inc., Customer Success / Data Analyst (Apr 2024 – Sep 2024)
  • Built Tableau dashboards from customer data, reducing reporting time by over 50% and supporting decision-making.
• Samsung Biologics, Market Intelligence Analyst Intern (Jun 2023 – Aug 2023)
  • Performed market and competitor analysis; automated recurring workflows, cutting manual effort by over 90%.
• Military Intelligence Service, 30th Mechanized Infantry Division (Mar 2019 – Oct 2020)
  • Analyzed intelligence data using ATCIS to support data-driven decisions (Corps Security Audit Commendation).

Research experience

• mHealth Team Communication Analysis, UCLA (Mar 2025 – present)
  • Analyzing 30+ multi-session team discussions with NLP for sentiment, role identification, and interaction dynamics.
  • Building Python pipelines for parsing, speaker diarization, and conversation-network metrics.
• SLIViT: Vision Transformer with Structured Local Interaction, UCLA (May 2025 – present)
  • Reproduced and verified SLIViT experiments; analyzed model structure and performance on visual tasks.

Publications

LLM Multi-Agent Simulations of Research Teams

Sungjin Choi. (2026). "LLM Multi-Agent Simulations of Research Teams." Under review, EMNLP 2026.

Revealing Team Project Development via Conversation Analysis

Sungjin Choi. (2025). "Revealing Team Project Development via Conversation Analysis." Working paper.

Skills

• Languages & tools: Python, SQL, R, Git, Docker, AWS, Google Cloud
• ML / DL: PyTorch, TensorFlow, scikit-learn
• Focus areas: LLM / LMM, NLP, deep learning, machine learning, image processing
• Analytics & viz: pandas, Tableau, statistical inference

Selected projects

• Soccer Player Market Value Scout, deployed Streamlit + FastAPI (Cloud Run) ML service. github.com/lacygus/418_final
• mHTI Multi-Agent Discussion Pipeline, M.S. thesis / EMNLP 2026 submission.
• Enhancing Emotion Recognition in AI, CLIP-Dissect interpretability. lacygus.github.io/Emotion_Recognition
• Congressional Stock-Trading Analysis, github.com/lacygus/Stocks-Trades

Contact

• Email: sungjinchoi5790@gmail.com
• GitHub: github.com/lacygus
• LinkedIn: sungjin-choi
• Location: Los Angeles, CA