Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I'm Desmond, a Generative AI and NLP specialist with over 5 years of data science experience. I have a strong passion for fine-tuning large language models, prompt engineering, and developing innovative GenAI applications using tools like LangChain and Hugging Face. My background includes managing token constraints, tuning model parameters, and automating testing pipelines to build reliable, enterprise-grade AI solutions. I have practical experience applying AI in various industries including healthcare and retail, working on projects from clinical data summarization to fraud detection. My education from UBC in deep learning and computational linguistics complements my hands-on skills, and I enjoy making complex AI technology accessible to both technical and non-technical audiences.…Hi, I'm Desmond, a Generative AI and NLP specialist with over 5 years of data science experience. I have a strong passion for fine-tuning large language models, prompt engineering, and developing innovative GenAI applications using tools like LangChain and Hugging Face. My background includes managing token constraints, tuning model parameters, and automating testing pipelines to build reliable, enterprise-grade AI solutions. I have practical experience applying AI in various industries including healthcare and retail, working on projects from clinical data summarization to fraud detection. My education from UBC in deep learning and computational linguistics complements my hands-on skills, and I enjoy making complex AI technology accessible to both technical and non-technical audiences.

Desmond (Tongyan) Bai

Data Scientist, AI Engineer, AI Developer, +2





Hi, I'm Desmond, a Generative AI and NLP specialist with over 5 years of data science experience. I have a strong passion for fine-tuning large language models, prompt engineering, and developing innovative GenAI applications using tools like LangChain and Hugging Face. My background includes managing token constraints, tuning model parameters, and automating testing pipelines to build reliable, enterprise-grade AI solutions. I have practical experience applying AI in various industries including healthcare and retail, working on projects from clinical data summarization to fraud detection. My education from UBC in deep learning and computational linguistics complements my hands-on skills, and I enjoy making complex AI technology accessible to both technical and non-technical audiences.…Hi, I'm Desmond, a Generative AI and NLP specialist with over 5 years of data science experience. I have a strong passion for fine-tuning large language models, prompt engineering, and developing innovative GenAI applications using tools like LangChain and Hugging Face. My background includes managing token constraints, tuning model parameters, and automating testing pipelines to build reliable, enterprise-grade AI solutions. I have practical experience applying AI in various industries including healthcare and retail, working on projects from clinical data summarization to fraud detection. My education from UBC in deep learning and computational linguistics complements my hands-on skills, and I enjoy making complex AI technology accessible to both technical and non-technical audiences.

Available to hire

Hi, I’m Desmond, a Generative AI and NLP specialist with over 5 years of data science experience. I have a strong passion for fine-tuning large language models, prompt engineering, and developing innovative GenAI applications using tools like LangChain and Hugging Face. My background includes managing token constraints, tuning model parameters, and automating testing pipelines to build reliable, enterprise-grade AI solutions.

I have practical experience applying AI in various industries including healthcare and retail, working on projects from clinical data summarization to fraud detection. My education from UBC in deep learning and computational linguistics complements my hands-on skills, and I enjoy making complex AI technology accessible to both technical and non-technical audiences.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Language

English

Fluent

Work Experience

Data Scientist Intern at Rocketbrew (Capstone Partner)

August 31, 2024 - August 1, 2025

Delivered an LLM-powered RAG solution to extract risk-related sentiment from customer verbatims using LangChain and GPT-4. Engineered custom prompt templates and chaining logic while adjusting temperature, top_p, and token window for better response specificity. Built prompt libraries and evaluation utilities enabling non-technical stakeholders to reliably interact with AI models. Integrated AI pipelines into Streamlit apps to support real-time fraud detection.

GenAI Systems Developer at DischargeMe Project

June 30, 2024 - August 1, 2025

Designed a two-part text-to-text summarization system using fine-tuned T5-small models to generate 'Brief Hospital Course' and 'Discharge Instructions' from over 100K raw EHR narrations. Tuned model parameters such as temperature and max token length to optimize fluency, factuality, and stylistic coherence. Built automated prompt testing and evaluation pipelines using metrics like ROUGE, BLEU-4, METEOR, BERTScore, and AlignScore. Managed hyperparameter tuning and input formatting for variable-length clinical data including handling inconsistent section headers. Deployed and tracked experiments using Hugging Face, Colab GPUs, and Weights & Biases and published trained models to Hugging Face Hub.

Senior Data Scientist at NielsenIQ

March 31, 2023 - August 1, 2025

Developed an anomaly detection pipeline and PyTorch-based fraud detection model for branded SKU datasets. Automated product bundling analytics for over 200 global clients and delivered insights via Docker-deployed reporting tools. Facilitated business alignment sessions to define AI output formats and embedded solutions into reporting frameworks.

Data Scientist at NielsenIQ

May 31, 2021 - August 1, 2025

Designed a ResNet classifier to clean visual SKU data, reducing image review time by 40%. Created Power BI dashboards and RShiny-based analytics tools to visualize COVID-19 impacts on supply chains. Implemented validation rules and anomaly detection logic for retail time-series datasets. Collaborated with engineering and data infrastructure teams to enhance model pipelines and streamline client reporting.