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

I'm Qing Dai, a full-stack AI engineer building cloud-deployed LLM-powered tools from data pipelines to production systems. I specialize in agentic retrieval-augmented generation workflows, model fine-tuning, and end-to-end deployment for real-world products. I collaborate across research and engineering to translate complex data into practical AI tooling. My recent work includes PSI IR benchmarking and the development of PSI's first agentic RAG chatbot, as well as AI-powered paper annotation tools at the University of Zurich.…I'm Qing Dai, a full-stack AI engineer building cloud-deployed LLM-powered tools from data pipelines to production systems. I specialize in agentic retrieval-augmented generation workflows, model fine-tuning, and end-to-end deployment for real-world products. I collaborate across research and engineering to translate complex data into practical AI tooling. My recent work includes PSI IR benchmarking and the development of PSI's first agentic RAG chatbot, as well as AI-powered paper annotation tools at the University of Zurich.

Qing Dai

AI Engineer, Full Stack Developer, Data Scientist, +2





I'm Qing Dai, a full-stack AI engineer building cloud-deployed LLM-powered tools from data pipelines to production systems. I specialize in agentic retrieval-augmented generation workflows, model fine-tuning, and end-to-end deployment for real-world products. I collaborate across research and engineering to translate complex data into practical AI tooling. My recent work includes PSI IR benchmarking and the development of PSI's first agentic RAG chatbot, as well as AI-powered paper annotation tools at the University of Zurich.…I'm Qing Dai, a full-stack AI engineer building cloud-deployed LLM-powered tools from data pipelines to production systems. I specialize in agentic retrieval-augmented generation workflows, model fine-tuning, and end-to-end deployment for real-world products. I collaborate across research and engineering to translate complex data into practical AI tooling. My recent work includes PSI IR benchmarking and the development of PSI's first agentic RAG chatbot, as well as AI-powered paper annotation tools at the University of Zurich.

Available to hire

I’m Qing Dai, a full-stack AI engineer building cloud-deployed LLM-powered tools from data pipelines to production systems. I specialize in agentic retrieval-augmented generation workflows, model fine-tuning, and end-to-end deployment for real-world products.

I collaborate across research and engineering to translate complex data into practical AI tooling. My recent work includes PSI IR benchmarking and the development of PSI’s first agentic RAG chatbot, as well as AI-powered paper annotation tools at the University of Zurich.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

German

Intermediate

Work Experience

Full-stack AI Engineer (Azure-Cloud & RAG, Part-time) at Xilva

October 1, 2025 - Present

Maintained and extended production cloud-deployed RAG web app for consultants, implementing embedding and vector-search pipelines and integrating Azure OpenAI models for document search, summarization, and analytics. Frontend: TypeScript/React; Backend: Python. Designed LLM routing and orchestration using Azure AI Foundry, managing prompt versions, model selection, latency, and cost monitoring across Azure-hosted services and Langfuse.

Master’s Thesis at Paul Scherrer Institute (PSI)

April 1, 2025 - November 30, 2025

Benchmarked a domain-specific framework for evaluating and optimizing Information Retrieval models in the accelerator-physics domain. Built the first expert-annotated IR dataset for domain model evaluation and automated annotation pipeline validated by seven domain experts; advised by Jingwei Ni (ETH AI Center) and Prof. Martin Volk (UZH).

Full-stack AI Engineer & Research Trainee on LLM at Paul Scherrer Institute PSI

November 1, 2024 - April 30, 2025

Designed and implemented PSI’s first agentic Retrieval-Augmented Generation (RAG) chatbot for technical document indexing and Q&A with source-linked answers. LangGraph orchestration; React/HTML/CSS frontend; Python backend; FastAPI REST API; LLM inference microservices; FAISS vector store; end-to-end workflow; LLM Evaluation; primary author of a paper presented at IPAC 2025.

Full-stack AI Engineer (AI-powered Paper Annotator) at DDIS Group, University of Zurich

May 1, 2024 - December 31, 2024

Integrated LLMs into an existing Chrome extension to automatically extract hypotheses and key components from scientific papers, linking them to a knowledge graph for richer queries and analytics. Conducted an A/B usability study with 16 researchers (NASA TLX) to measure task time and cognitive load.

Software Engineer (Recommender System) at University of Zurich

September 1, 2023 - January 31, 2024

Built a movie chatbot with a recommender system using item-to-item similarity with vector embeddings. Implemented NLP data pipeline (spaCy NER and Intent Classification) and integrated Wikidata knowledge-graph queries to improve entity resolution and outputs.

Senior Supply Chain Management Planner at Axalta (DuPont) Coating System

April 1, 2018 - June 1, 2021

Drove supply strategies for multiple business segments; aligned demand with production capacity; monitored production, region inventory, and shipments; analyzed historical sales data and forecasts; developed contingency plans to minimize disruption.

Customer Service & Master Planner at Magna Interiors

April 1, 2012 - March 1, 2018

Coordinated requirements from major automotive clients; led cross-functional teams to meet customer needs; managed production planning for on-time delivery; SAP Production Module maintenance.

Full-stack AI Engineer (Azure-Cloud & RAG, Part-time Work) at Xilva

October 1, 2025 - Present

Maintained and extended a production, cloud-deployed RAG web application for consultants, implementing embedding and vector-search pipelines and integrating Azure OpenAI models for document search, summarization, and analytics. (Typescript/React frontend, Python backend). Designed LLM routing and orchestration using Azure AI Foundry, managing prompt versions, model selection, latency, and cost monitoring across Azure-hosted services and Langfuse.

Master’s Thesis Researcher at Paul Scherrer Institute (PSI)

April 1, 2025 - November 1, 2025

Benchmarked a domain-specific Information Retrieval framework for evaluating and optimizing IR models in the accelerator-physics domain. Created the first expert-annotated IR dataset as a gold-standard for IR model evaluation; developed an automated annotation pipeline; advised by Jingwei Ni (ETH AI Center) and Prof. Martin Volk (UZH).

Full-stack AI Engineer & Research Trainee on LLM at Paul Scherrer Institute (PSI)

November 1, 2024 - April 1, 2025

Designed and implemented PSI’s first agentic Retrieval-Augmented Generation (RAG) chatbot for technical document indexing and Q&A with source-linked answers. LangGraph orchestration, React/HTML/CSS frontend, Python backend, FastAPI REST API, LLM inference microservices, FAISS vector store, end-to-end workflow; assessed models as a judge and evaluated them.