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

I’m an experienced developer specializing in RAG systems, multi-agent AI, AI applications, and end-to-end machine learning solutions. Currently at LexisNexis, I design evaluation pipelines, build agentic search and evidence-tracing tools, and develop LLM-powered systems for legal intelligence. Previously, I worked on NLP, computer vision, voice conversion, and digital human projects in fintech, with experience taking AI models from research to production. I hold a Master’s degree in Engineering with a Data Analytics emphasis from the University of Toronto, and I’m passionate about building reliable, scalable AI systems that solve real-world problems.…I’m an experienced developer specializing in RAG systems, multi-agent AI, AI applications, and end-to-end machine learning solutions. Currently at LexisNexis, I design evaluation pipelines, build agentic search and evidence-tracing tools, and develop LLM-powered systems for legal intelligence. Previously, I worked on NLP, computer vision, voice conversion, and digital human projects in fintech, with experience taking AI models from research to production. I hold a Master’s degree in Engineering with a Data Analytics emphasis from the University of Toronto, and I’m passionate about building reliable, scalable AI systems that solve real-world problems.

Lin Qizhi

Data Scientist, AI Engineer, Data Analyst, +2





I’m an experienced developer specializing in RAG systems, multi-agent AI, AI applications, and end-to-end machine learning solutions. Currently at LexisNexis, I design evaluation pipelines, build agentic search and evidence-tracing tools, and develop LLM-powered systems for legal intelligence. Previously, I worked on NLP, computer vision, voice conversion, and digital human projects in fintech, with experience taking AI models from research to production. I hold a Master’s degree in Engineering with a Data Analytics emphasis from the University of Toronto, and I’m passionate about building reliable, scalable AI systems that solve real-world problems.…I’m an experienced developer specializing in RAG systems, multi-agent AI, AI applications, and end-to-end machine learning solutions. Currently at LexisNexis, I design evaluation pipelines, build agentic search and evidence-tracing tools, and develop LLM-powered systems for legal intelligence. Previously, I worked on NLP, computer vision, voice conversion, and digital human projects in fintech, with experience taking AI models from research to production. I hold a Master’s degree in Engineering with a Data Analytics emphasis from the University of Toronto, and I’m passionate about building reliable, scalable AI systems that solve real-world problems.

Available to hire

I’m an experienced developer specializing in RAG systems, multi-agent AI, AI applications, and end-to-end machine learning solutions. Currently at LexisNexis, I design evaluation pipelines, build agentic search and evidence-tracing tools, and develop LLM-powered systems for legal intelligence. Previously, I worked on NLP, computer vision, voice conversion, and digital human projects in fintech, with experience taking AI models from research to production. I hold a Master’s degree in Engineering with a Data Analytics emphasis from the University of Toronto, and I’m passionate about building reliable, scalable AI systems that solve real-world problems.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

Senior Data Scientist / LLM Engineer at LexisNexis Shanghai

May 1, 2024 - Present

Led the design and development of end-to-end evaluation pipelines for legal RAG and AI agent systems, with tailored rating tools and multi-dimensional metrics to assess intermediate and final outputs. Contributed to a multi-agent system, independently building the core module Doc Locator to trace sentence-level evidence and enable precise source referencing. Built a universal legal search tool for agent systems featuring multi-source hybrid search, asynchronous multi-process retrieval, results reflection, and context management. Orchestrated an automated prompt optimization pipeline using the GEPA framework. Maintained and upgraded the legal RAG system, contributing to modules including intent detection, query decomposition, and vector search tools. Fine-tuned an Mistral-7B LLM with LoRA for intent detection, achieving 96% accuracy. Developed an automatic AI topic recommendation platform to discover, filter, summarize, and grade AI news from diverse sources and publish daily topics.

Algorithm Engineer at Lufax Holding Ltd.

June 1, 2021 - May 1, 2024

Developed a ChatGLM-based verifier using LoRA and fastllm for credit approval scenarios, enabling real-time customer qualification analysis and dialogue evaluation training. Served as project manager for a voice conversion model, delivering R&D and production with streaming architecture and end-to-end latency under 500ms. Built a CV system covering face detection, facial key points, alignment, recognition, and image quality detection, deployed in production for risk control. Contributed to a digital human project by integrating a wav2lip-based generation model with a streaming voice conversion model to guide customers through processes. Implemented an intent recognition model based on SimBERT and Elasticsearch to support a real-time QA chatbot. Led a team to compete in multiple SemEval 2022 tracks, securing first place in Idiomaticity Detection.

Data Scientist Intern at TasteGuru.ai

May 1, 2019 - December 1, 2019

Developed a Dynamic Topic Model on social media data to capture historical trends in the beer industry, clustering millions of tweets. Built an automated web scraping script to collect data for the database and performed data analysis and visualization on Google Cloud Platform and AWS.