☀️
Kailong Duan
Full-Stack Software Engineer · AI/ML Enthusiast
Results-oriented Full-Stack Software Engineer with 5 years of professional experience designing, developing, and deploying scalable enterprise applications. Proficient in Vue.js, React.js, TypeScript and Java / Spring Boot and Python / Django / FastAPI backends. Skilled in deploying full-stack applications on AWS and working across the full development lifecycle. Currently expanding expertise in AI/ML and NLP using Python and LLM-based models. Seeking roles starting May 2026.
Software Engineer Intern
📍 Dallas, TX
- Built an internal React dashboard for real-time semiconductor test data visualization, consolidating data from 5+ engineering teams and reducing manual QA review time by ~30%.
- Developed RESTful APIs using Python / FastAPI and PostgreSQL to expose chip testing metrics; cut average data retrieval latency from 200ms to 80ms through query optimization and Redis caching.
- Automated test report generation with Python and Docker, processing 10,000+ daily test records and deploying as a scheduled AWS Lambda function — eliminating 4+ hours of manual reporting per week.
- Collaborated with firmware and hardware validation teams to streamline CI/CD workflows via Jenkins and GitLab CI, reducing build-to-deploy cycle time by 25% for internal tooling releases.
Software Engineer
📍 Beijing, China
- Architected and maintained Spring Boot microservices with MySQL and Kafka, supporting core business workflows for 10,000+ daily active users across enterprise clients.
- Designed RESTful APIs with Docker-based packaging, enabling reliable frontend-backend communication and reducing integration bugs by ~40%.
- Improved system throughput by 35% through query optimization, indexing strategies, and a Redis caching layer that cut average DB response time from 120ms to 45ms.
- Implemented secure JWT + Spring Security (RBAC) authentication, eliminating unauthorized access incidents across all services.
Software Engineer
📍 Shenzhen, China
- Built mission-critical SPAs and mobile features with Vue.js, React.js, and TypeScript, serving 500+ concurrent users with sub-2s page load times.
- Deployed frontend services via Nginx with load balancing and caching, reducing peak-hour latency by 30% and improving system availability to 99.9%.
- Established CI/CD pipelines using Jenkins and GitLab, cutting deployment lead time by 35% and eliminating manual release errors.
M.S. in Software Engineering
📍 Dallas, TX
- Built and deployed a cloud-based application on AWS — React frontend via CloudFront & S3, MySQL backend on AWS RDS.
- Developed backend management systems using Django and Oracle as part of coursework and self-directed projects.
- Built and trained ML models and reinforcement learning agents (Q-learning, DDPG, PPO) using TensorFlow, PyTorch, and OpenAI Gym. Currently researching RAGFlow for retrieval-augmented LLM workflows.
B.S. in Computer Science
📍 Zhengzhou, China
PythonPyTorchCNNDockerFlaskPostgreSQL
- Architected and trained a custom ResNet-18 model from scratch on CIFAR-10, achieving over 92% Top-1 accuracy.
- Implemented industrial-grade training pipelines with AMP, Cosine Annealing with Linear Warmup, Cutout augmentation, and Label Smoothing to enhance generalization and accelerate convergence.
- Developed interpretability modules using Grad-CAM to visualize neural network activation heatmaps and diagnose model behavior.
- Built and deployed an end-to-end Flask web app serving the trained model with real-time inference and probability distribution visualizations.
PythonDocker ComposeOllamaRAGLLMsTensorFlowEmbeddings
- Deployed a comprehensive RAG pipeline capable of ingesting heterogeneous unstructured data using deep document understanding for accurate knowledge retrieval.
- Integrated custom local LLM runners (Ollama) with the RAG framework to build a privacy-first grounded generation system.
- Managed intelligent document chunking and embedding, reducing LLM hallucinations by ensuring highly relevant context and traceable citations.
- Configured advanced recall mechanisms and seamless API interactions to bridge document vectorization and final text synthesis.
ReactVitePythonMySQLTailwind CSSRESTful APIs
- Engineered a full-stack web application with a React/Vite SPA frontend and Python/MySQL backend supporting complex social networking and research data workflows.
- Normalized relational database schemas in MySQL from ER models covering Users, Posts, Projects, and Institutes; built scalable RESTful APIs for multi-parameter queries and statistical aggregation.
- Implemented advanced SQL aggregation logic and real-time data pipelines delivering low-latency JSON responses to React component state.
React 19TypeScriptViteSpring BootKafkaSwaggerMUI
- Built an end-to-end productivity system with a React SPA frontend and a Java / Spring Boot backend, covering calendar scheduling, task workflows, timer sessions, user profile, and authentication.
- Implemented a centralized API service layer for task and auth operations (task CRUD, completion updates, user/profile/auth requests), improving maintainability and reducing page-level coupling.
- Developed Pomodoro / short-break / long-break workflows with reliable start/pause/reset behavior using React Hooks for interval lifecycle control.
- Delivered backend APIs with Spring Boot + Swagger and introduced Kafka-based event handling for task-notification flows to support scalable decoupled architecture.
Languages
JavaPythonJavaScriptTypeScriptSQLHTML/CSS
Frameworks
Spring BootReact.jsVue.jsDjangoFastAPIFlaskTailwind
Databases & MQ
MySQLPostgreSQLOracleRedisApache Kafka
Cloud & DevOps
DockerJenkinsGitLab CI/CDNginxLinuxGit
Cloud (AWS)
EC2S3CloudFrontRDSLambda
AI / ML
TensorFlowPyTorchLLM IntegrationRAGVector DBOpenAI APIPPO