AI Application Engineer
Hi, I'm Guoqing Liu
AI Application Engineer focused on speech recognition, LLM API development, and deploying multimodal AI solutions.
01.About Me
AI Application Engineer at Beijing FLTRP Online Digital Technology. Responsible for R&D across LLM applications, speech technology, digital human video generation, and intelligent scoring systems — independently delivering multiple projects from 0 to 1.
Specialized in turning AI into production systems: from fine-tuning Whisper for pronunciation scoring, to building a fact-checking engine (90.6% precision), to a microservice-based digital human platform (75+ hours of video generated). Passionate about pushing AI boundaries and turning complex model capabilities into tangible product value.
03.Projects
Oral English Auto-Scoring
ASR + Rule Engine + LLM multi-layer pipeline for automated oral exam scoring with speech transcription and intelligent grading.
Fact-Checking Engine
LLM + multi-source evidence retrieval + 11 modular Checkers for automated fact verification. Dual-parallel architecture, 90.6% precision, 1.5% FPR.
AI Digital Human Courseware
Microservice-based end-to-end video production pipeline. PPT/docs → digital human videos. 24 Pod parallel GPU inference, 75+ hours generated.
Pronunciation Scoring Model
Whisper large-v3 based scoring system with Cross-Attention + Attention Pooling architecture, trained on 150K labeled samples.