Deng Yulin

Senior AI Application Engineer

📧 1016068291@qq.com | 📱 (+86) 17679329004 | 📍 Shanghai, China | 7 Years Experience

Technical Skills

AI/Machine Learning Tech Stack

  • LLM Application Development: LlamaIndex, LangChain, Transformers, OpenAI API, Claude API
  • Vector Databases: Weaviate, FAISS, Pinecone, Qdrant, Chroma
  • RAG Systems: Document parsing, vectorization, semantic retrieval, answer generation, context optimization
  • Prompt Engineering: Prompt optimization, Few-shot Learning, Chain-of-Thought, template design
  • Model Optimization: Fine-tuning, LoRA, quantization deployment, inference optimization
  • AI Frameworks: PyTorch, TensorFlow, Hugging Face, scikit-learn

Backend Development & Architecture

  • Programming: Python (8 years), concurrent programming, asynchronous programming, network communication
  • Web Frameworks: FastAPI, Django, Flask, Celery, DRF
  • Databases: MySQL (index optimization, transaction management, sharding), PostgreSQL, MongoDB
  • Caching: Redis (clustering, persistence, cache strategy optimization)
  • Message Queues: RabbitMQ, RocketMQ, Kafka

Cloud Native & DevOps

  • Containerization: Docker, Kubernetes, microservices architecture design
  • Cloud Platforms: Alibaba Cloud, AWS, Tencent Cloud
  • Monitoring: Prometheus, Grafana, ELK Stack
  • CI/CD: Jenkins, GitLab CI, automated deployment

Professional Experience

Shanghai JIDU Automotive Co., Ltd. - Senior AI Application Engineer

Jan 2022 - Dec 2024

AI-Powered Document Security Inspection System

  • Built document content understanding and risk identification system using Large Language Models, combined with NLP techniques for automatic document classification
  • Designed vector similarity-based document risk detection algorithms with 95%+ accuracy, reducing high-risk files from 30,000+ to 0
  • Implemented document parsing pipeline using LangChain, supporting intelligent content extraction and analysis for multiple formats (PDF/Word/Excel)

JAMA Intelligent Requirements Management Platform

  • Developed GPT-based intelligent requirements document review system with automatic completeness and consistency detection
  • Implemented dynamic permission allocation algorithm using historical data-trained prediction models, saving $300K+ annually
  • Integrated Prompt Engineering techniques to enable automatic generation and standardization of requirements documents

Enterprise AI Toolchain Integration

  • Designed and developed 2000+ automation scripts integrating AI capabilities across Jira, Feishu, and WeChat Work platforms
  • Implemented NLP-based intelligent ticket classification and routing system, improving processing efficiency by 300%+
  • Built intelligent workflow for cross-system data synchronization with automatic adaptation to organizational changes

Shanghai Data Exchange Co., Ltd. - AI Application Engineer

Feb 2020 - Jan 2021

Intelligent Data Trading Recommendation System

  • Designed and implemented hybrid recommendation algorithms combining collaborative filtering and content-based approaches, improving user conversion rate by 40%
  • Optimized data matching accuracy using deep learning models through feature engineering and model tuning, achieving 30% performance improvement
  • Built real-time recommendation service supporting high-concurrency scenarios with millisecond-level response

Backend Service Performance Optimization

  • Enhanced system processing capability by 200% and reduced response time by 50% through asynchronous processing and caching strategies
  • Led database query optimization combining indexing strategies and SQL tuning, improving query efficiency by 50%
  • Implemented service monitoring and auto-scaling mechanisms, achieving 99.9% system availability

Shanghai Yilu Investment Management Co., Ltd. - Python Engineer

Feb 2018 - Jan 2020

Smart Hotel AIoT Platform

  • Led smart hotel hardware product selection and integration, designed hardware communication protocols and low-level encapsulation
  • Developed facial recognition system integrating computer vision algorithms for customer identity verification and personalized services
  • Built hotel PMS system with integrated AI capabilities including intelligent customer service, dynamic pricing, and demand forecasting

Key AI Projects

Enterprise RAG Knowledge Q&A System

Jun 2023 - Nov 2024 | Technical Lead

Background: Built enterprise internal knowledge base intelligent Q&A system for JIDU Automotive, supporting multi-source data retrieval including technical documents, standards, and historical projects

Technical Implementation:

  • Built multimodal document parsing pipeline using LlamaIndex, supporting PDF, Word, PPT formats
  • Implemented Chinese document vectorization using BGE models with Weaviate as vector database
  • Designed hybrid retrieval strategy (semantic + keyword search), improving retrieval accuracy by 35%
  • Integrated ChatGPT/Claude APIs for context-aware answer generation
  • Built Prompt template library with customized response strategies for different business scenarios

Key Results: Deployed to production serving 1000+ employees, 5000+ daily queries, 90%+ satisfaction rate

Intelligent Ticket Classification & Processing System

Aug 2022 - May 2023 | Core Developer

Background: Implemented intelligent Jira ticket classification, priority prediction, and automatic assignment using NLP technologies

Technical Implementation:

  • Used BERT model for ticket text classification with 92% accuracy
  • Built priority prediction model based on historical data combined with time series analysis
  • Implemented intelligent routing algorithm based on engineer skill profiles and workload
  • Integrated Slack/Feishu bots for intelligent ticket status notifications and reminders

Key Results: 60% improvement in ticket processing efficiency, 40% reduction in average resolution time

Multimodal Content Safety Detection Platform

Mar 2022 - Dec 2022 | Technical Lead

Background: Built enterprise content safety intelligent review system supporting security risk identification for text, images, and documents

Technical Implementation:

  • Integrated multiple pre-trained models: text classification (RoBERTa), image recognition (ResNet), OCR (PaddleOCR)
  • Designed hybrid detection strategy combining rule engine and machine learning
  • Built real-time detection pipeline supporting API calls and batch processing
  • Implemented explainable analysis providing risk point annotation and confidence scoring

Key Results: 95%+ detection accuracy, false positive rate controlled under 5%, processing 1M+ content daily

Technical Highlights

AI System Architecture Design

  • Complete AI application development lifecycle management experience from requirement analysis to model deployment and operations
  • Expert in designing high-concurrency, low-latency AI service architectures supporting tens of millions of users
  • Proficient in model service deployment including quantization, inference optimization, and A/B testing

Cross-Domain Technology Integration

  • Deep integration of AI capabilities with traditional business systems for intelligent transformation
  • Rich experience in enterprise system integration, skilled in extending AI capabilities for Atlassian, Feishu, WeChat Work platforms

Data Processing & Optimization

  • Deep understanding of vector database principles, expert in performance tuning for large-scale vector retrieval systems
  • Experienced in ETL pipeline design supporting data cleaning, transformation, and vectorization for multi-source heterogeneous data

Education

Jiangxi University of Science and Technology | Electronic Information Engineering | Bachelor’s Degree | 2013-2017

Relevant Coursework: Digital Signal Processing, Machine Learning Fundamentals, Data Structures & Algorithms, Software Engineering

Professional Strengths

✅ 7 years Python development + 3 years AI application development experience
✅ End-to-end experience building enterprise-level AI applications from scratch
✅ Deep understanding of LLM application architecture design and performance optimization
✅ Rich cross-functional collaboration experience, able to deeply integrate AI technology with business requirements
✅ Continuous learning of latest AI technology trends with rapid technology transfer and innovation capabilities

__END__