Terry Li

Education

Publications & manuscripts

1. Accepted

   Peilin Li. Edge-Intelligence-Enhanced Proximal Policy Optimization (EIF-PPO) for Real-Time Power Balancing in Distributed Energy Microgrids. Informatica.

2. Accepted

   AI-Driven Lung Cancer Screening: A Comparative Analysis of Machine Learning Models. 2024 International Conference on Modern Logistics and Supply Chain Management.

3. In preparation

   Peilin Li. Application of Deep Reinforcement Learning in Optimizing Virtualized GPU Resource Scheduling for Real-time Rendering. Undergraduate Research Program, Tsinghua University.

4. In preparation

   Peilin Li. SPEM: A Split-Timescale Coupled Error Model for Stochastic-Photonic GEMM. Undergraduate Research Program, Tsinghua University.

1. Aug 2026 – Expected May 2027 Research

   Researcher · PURE-ECE Undergraduate Research Program, University of Michigan

   Universal Operations for Compositional Design

   Advisor: Prof. Inigo Incer · Ann Arbor, MI

   DetailsHide details ›

   • Developing an automated framework for studying universal assume-guarantee contract operators, replacing error-prone manual evaluation of three-valued logic across nested compositions.
   • Implementing a recursive abstract-syntax-tree evaluator that enumerates ternary inputs, generates truth tables, and analyzes operator properties and behavioral equivalence across configurable expression depths.
   • Building diagrams and animations to trace composition structures and logical transitions, using weekly research reviews to diagnose anomalous cases, validate computational results, and refine evaluation methods.

2. Jul 2025 – Expected Jan 2027 Research

   Researcher · Undergraduate Research Program, Tsinghua University

   Reinforcement Learning-Based Adaptive Scheduling for Virtualized GPU Systems

   Advisor: Prof. Yiyang

   DetailsHide details ›

   • Investigating deep reinforcement learning for virtualized GPU scheduling, addressing resource underutilization and latency instability caused by time-varying real-time rendering workloads; formulating adaptive resource allocation as a sequential decision-making problem.
   • Developing a closed-loop scheduler that converts VM/GPU telemetry into workload states and dynamically selects GPU partitioning, preemption, and priority actions, with a reward function balancing utilization, task latency, service quality, and reconfiguration overhead.
   • Evaluating the learned policy against static partitioning and fixed-priority baselines under changing execution conditions using GPU utilization, tail latency, SLA violations, and scheduling overhead; preparing an EI-indexed journal manuscript on the resulting framework and experimental findings.

3. Aug 2026 – Expected Dec 2026 Research

   Researcher · Lab of Geometry at Michigan LoG(M), University of Michigan

   A sphere by any other name

   Advisor: Dr. Neza Zager Korenjak · Ann Arbor, MI

   DetailsHide details ›

   • Investigating how alternative bilinear forms change the geometry of generalized spheres, with emphasis on symmetries, embedded circles, positivity, and cyclic order.
   • Constructing and comparing low-dimensional examples using matrix representations and linear-algebraic analysis to identify invariant and form-dependent geometric properties.
   • Developing computational visualizations and geometric diagrams that make abstract boundary-order structures interpretable and support further theoretical analysis.

4. Feb 2026 – Expected Dec 2026 Research

   Researcher · Undergraduate Research Program, Tsinghua University

   SPEM: A Split-Timescale Coupled Error Model for Stochastic-Photonic GEMM

   Advisor: Prof. Yiyang

   DetailsHide details ›

   • Developed SPEM, a split-timescale model coupling stochastic-sampling variance with microring drift, shot noise, and WDM crosstalk to predict irreducible noise floors and bitstream saturation in stochastic-photonic GEMM.
   • Built a disjoint calibration and Monte Carlo validation pipeline mapping measurable hardware parameters to GEMM-level NFMSE.
   • Achieved 1.96% median prediction error over 200 held-out configurations and exact optima on a 234-point design grid; reduced sampling energy by up to 97.8% under matched error budgets and derived thermal-drift guard bands for a first-author journal manuscript.

5. May 2026 – Aug 2026 Internship

   Engineer Intern · Artisk INC

   Backend services and APIs, data models, cloud authentication and file-storage integration, automated tests.

   San Francisco, CA

   DetailsHide details ›

   • Developed and maintained backend services and APIs to support core product workflows.
   • Designed data models, performed database operations, and integrated cloud services including authentication and file storage.
   • Troubleshot system issues, authored automated tests, and optimized the reliability and performance of existing systems.

6. Jul 2025 – Jun 2026 Research

   Sole Author · University of Illinois Urbana-Champaign

   Power Balancing Control Method for Distributed Energy Microgrids Based on Edge Computing — Informatica (forthcoming 2026)

   Champaign, IL

   DetailsHide details ›

   • Formulated real-time microgrid power balancing as a sequential decision problem and developed EIF-PPO to address power mismatch, frequency deviation, and battery instability under renewable intermittency and cloud-communication delays.
   • Deployed a PPO actor–critic controller within a hierarchical edge architecture, transforming asynchronous PV, BESS, and load-bus telemetry into decentralized dispatch actions; constructed LMOD-based stochastic environments to emulate generation, demand, and communication uncertainty.
   • Achieved a 62.5% reduction in inference latency and a 41% improvement in battery SoC stability relative to cloud-based control, independently completing framework design, simulation, benchmarking, analysis, and manuscript development.

7. Jul 2024 – Sep 2024 Research

   Researcher · Summer Research Program, Carnegie Mellon University

   AI-Driven Lung Cancer Screening: A Comparative Analysis of Machine Learning Models

   DetailsHide details ›

   • Conducted a comparative study of machine-learning methods for lung cancer screening using Kaggle datasets, examining differences in predictive performance and high-risk-case detection across classification approaches.
   • Implemented a reproducible Python and TensorFlow/Keras workflow for data preprocessing, model training, and held-out testing, evaluating classifiers using accuracy, precision, recall, F1-score, ROC-AUC, and confusion-matrix analysis.
   • Synthesized model comparisons and error-analysis findings into a co-authored research paper accepted at the 2024 International Conference on Modern Logistics and Supply Chain Management.

8. 2024 Award

   Platinum Division · USA Computing Olympiad (USACO)

9. 2024 Award

   National Gold · 2024 Mathematics Competition

10. 2024 Award

    National Gold · 2024 PHYSICS BOWL

11. 2024 Award

    VEX V5 Student Certification · VEX Robotics

12. 2023 Award

    3rd Place · 2023 World Robot Contest Finals

13. 2023 Award

    Youth Creativity Invention Award · Simulated Traffic Lights (STM32 prototype)

14. — Award

    Silver Medal (Top 4%) · Kaggle — HMS: Harmful Brain Activity Classification

Technical skills

Programming

C++JavaPythonSQLJavaScript

Machine Learning & AI

TensorFlowKerasOpenAI APIRAG ArchitecturePineconeElasticsearch

Backend & Systems

FPGA: SystemVerilogSpring BootMicroservicesRedisREST APIsJWT

Cloud & Infrastructure

AWS (ECS, Lambda, SQS, S3)Kubernetes (EKS)DockerGoogle Cloud Run

Computer Science Foundations

Data Structures & Algorithms (USACO Platinum)Operating Systems