Teaching

Teaching philosophy, experience, mentoring, and course interests.

Teaching and mentoring let me contribute to the progress of future engineers and researchers. My goal is to help students connect mathematical principles with implementation, choose appropriate methods and tools, and evaluate results against explicit requirements.

Teaching Philosophy

I organize my teaching around three connected practices:

  • Understanding: students explain a method’s assumptions, expected behavior, and suitable problem settings.
  • Implementation: students translate mathematical definitions into code and compare theory with observed behavior.
  • Validation: students state requirements, inspect evidence, and explain what a test or formal check establishes and what it leaves unresolved.

In Learning and Planning in Robotics, I used guided discussion to compare Q-learning and dynamic programming through their assumptions and appropriate settings. In Linear Algebra and Component Analysis, I connected hand calculations, nested-loop matrix multiplication, NumPy, Strassen’s algorithm, and timing experiments. Students then selected a method for a stated setting and defended the choice using mathematics and observed behavior.

My Vanderbilt guest lectures apply the validation practice to AI-assisted engineering. In Cyber-Physical Artificial Intelligence, students examined a workflow from specifications to acceptance and discussed what evidence is needed beyond an agent’s claim that a task is complete. Across three survey items about that lecture, 52 of 55 item responses (94.5%) indicated agreement or strong agreement; these are pooled item responses, not 55 distinct students. In Formal Methods: Safe and Trustworthy AI, my segment used a file-deletion example to show how a model and runtime monitor can enforce human approval as a precondition.

Teaching Experience

Vanderbilt University

  • CS 3892/5892, Formal Methods: Safe and Trustworthy AI, Fall 2026 — Guest Lecturer; specifications and runtime enforcement
  • CS 6378, Cyber-Physical Artificial Intelligence, Fall 2026 — Guest Lecturer; human–agent collaboration from specifications to acceptance
  • CS 6315, Automated Verification, Spring 2026 — Guest Lecturer
  • CS 3892/5892, Projects in Cloud Computing, Fall 2025 — Guest Lecturer

Washington University in St. Louis

  • ESE 559, Learning and Planning in Robotics, Spring 2025 — Assistant Instructor
  • ESE 2180, Linear Algebra and Component Analysis, Fall 2024 — Leading Assistant Instructor

Mentoring

My mentoring connects mathematical understanding with implementation and scientific insight. Students examine assumptions, identify principles behind observed behavior, propose methods, and validate claims through analysis or experiments.

I have mentored thirteen undergraduate and graduate researchers:

  • Vanderbilt graduate directed research: Yilin Liu and Guocheng He
  • Vanderbilt undergraduate directed research: Aarij Atiq
  • Washington University Research Experience for Undergraduates: Ziyi Peng, Mario Rodriguez-Montoya, Dzenan Zecevic, Nicolas Hernandez, and Keyron Linarez
  • Washington University graduate directed research: Mark Charnot and Seifeldin I. Elkhashab
  • Worcester Polytechnic Institute Major Qualifying Projects: Noelle Johnson, Minh Le, and Christopher Letherbarrow

Teaching Interests

I am prepared to teach Engineering Mathematics, Signals and Systems, Probability and Stochastic Processes, Introduction to Optimization, Introduction to Control, Linear Systems, and Embedded Control Systems. At advanced undergraduate and graduate levels, I am interested in Optimal Control, Formal Verification, Introduction to Robotics, and Robotic Planning.

I also propose an elective, Trustworthy AI-Enabled Cyber-Physical Systems. The course would cover formal specifications, control barrier functions and safety filters, neural-network verification, and runtime monitoring. A proposed course project would ask students to build a controller for a simulated system with an AI coding agent, state a safety requirement, and implement a safety filter or runtime monitor. Assessment would emphasize what the student’s evidence establishes and what remains unverified. This is a proposed course, not a previously delivered course.