Li Jiayin

Led the development and deployment of advanced GenAI and NLP solutions, focusing on transformer-based models and scalable microservices for production environments.

• Led development of multilingual NLP pipelines using HuggingFace Transformers for complex NER, classification, and summarization tasks.
• Engineered scalable FastAPI microservices, deploying them efficiently via Docker and AWS ECS to support high-performance AI applications.
• Integrated Large Language Models (LLMs) into QA workflows with human-in-the-loop tools, boosting annotation accuracy by 30%.
• Optimized model serving by deploying transformer models via Triton Inference Server with ONNX, significantly improving inference performance.
• Orchestrated high-volume asynchronous workloads using Celery and Redis, successfully supporting over 1 million monthly API requests.

Designed and implemented robust backend services and AI applications, focusing on data pipelines, real-time monitoring, and production model integration for enterprise clients.

• Designed robust backend services for enterprise AI applications, leveraging Python and FastAPI for high-performance and scalability.
• Developed critical predictive maintenance tools and document understanding pipelines, delivering specialized solutions for government and finance clients.
• Deployed real-time monitoring dashboards and data APIs using PostgreSQL and Redis, enhancing operational visibility and data accessibility.
• Integrated Scikit-learn and XGBoost models into production-grade services, ensuring reliable performance via REST and gRPC APIs.
• Led the successful rollout of containerization and CI/CD pipelines using GitHub Actions and Docker on AWS, significantly streamlining deployment processes.

Developed educational backend systems and introduced modern software development practices while instructing undergraduate students in computer science.

• Developed Flask-based backend systems to automate student progress tracking and grading, improving administrative efficiency for faculty.
• Introduced Git-based project workflows and DevOps practices into the Computer Science curriculum, enhancing student development methodologies.
• Engineered Firebase Auth-based portals with role-based multilingual content delivery, expanding access and usability for diverse student populations.
• Delivered comprehensive instruction in data science, algorithms, and practical Python programming to undergraduate students, fostering foundational skills.

Engineered educational technology tools and digitized administrative processes while teaching mathematics and leading curriculum modernization initiatives.

• Developed educational tools using C++ and Python to facilitate computational math and problem-solving skills for students.
• Digitized key administrative workflows through custom SQL-backed internal tools, enhancing operational efficiency and data management.
• Designed assistive learning modules and UI components tailored for special education use cases, improving accessibility and engagement.
• Led a successful university-wide initiative to integrate Python, Git, and data science into engineering programs, modernizing the curriculum.

A GenAI-powered trading bot designed to generate strategies and execute trades in the cryptocurrency market.

A web application developed to automate conflict-free scheduling for academic timetables.

An automated pipeline for preprocessing large datasets of panoramic images for machine learning applications.