CV

Curriculum vitae for Varun Rayamajhi, an undergraduate robotics researcher working on multi-agent systems, reinforcement learning, and safety-critical control.

Contact Information

Name Varun Rayamajhi
Professional Title Undergraduate Student
Email varun [dot] rayamajhi [at] richmond [dot] edu

Professional Summary

Undergraduate researcher interested in multi-agent robotics, reinforcement learning, and safety-critical control.

Experience

  • 2026 - Present

    Pittsburgh, PA, USA

    Undergraduate Research Fellow
    Carnegie Mellon University, Robotics Institute
    Advisor: Prof. Katia Sycara, Advanced Agent-Robotics Technology Lab
    • Developing safe hierarchical multi-agent reinforcement learning methods for persistent environmental sensing in the presence of spatiotemporal risks such as fire and smoke.
  • 2025 - Present

    Richmond, VA, USA

    Undergraduate Researcher
    University of Richmond, Robotics Lab
    Advisor: Dr. Patrick Martin
    • Investigating whether evolving Adaptive Control Barrier Function penalty variables can forecast rising safety stress before a robot reaches a safety-critical state.
  • 2025 - 2025

    Chicago, IL, USA

    Research Software Engineering Intern
    University of Chicago, Data Science Institute Summer Lab
    Advisor: Dr. Pedro Lopes, Human-Computer Integration Lab
    • Developed the first interactive mobile application for systematic digital measurement of downgaze palsy in Progressive Supranuclear Palsy assessment.
    • Reduced evaluation time by 80% compared with the traditional manual workflow in the project study.
    • Developed a three-stage computer-vision pipeline for automating downgaze palsy assessment by combining image processing, machine-learning techniques, and geometric analysis.
  • 2025 - 2025

    Richmond, VA, USA

    Machine Learning Research Assistant
    University of Richmond, Department of Physics
    Advisor: Dr. Jack Singal
    • Investigated Lagrangian and Hamiltonian Neural Networks for a dissipative mechanical system with explicit time dependence in its Lagrangian, Hamiltonian, and total energy.
    • Built an end-to-end PyTorch pipeline for synthetic data generation, preprocessing, training, hyperparameter tuning, validation, and performance analysis.
    • Presented the project at the Edinburgh x Anthropic AI Expo 2026 after selection as one of 25 projects from 70 submissions.
  • 2024 - 2024

    Richmond, VA, USA

    Undergraduate Robotics Researcher
    University of Richmond, Robotics Lab
    Advisor: Dr. Patrick Martin
    • Developed a ROS 2 framework for decentralized multi-agent formation control and leader-follower coordination.
    • Implemented action servers, clients, custom messages, and launch files for scalable agent deployment and communication.
    • Validated the controllers in simulation and on a three-TurtleBot physical system.
  • 2026 - Present

    Richmond, VA, USA

    Founder and President
    Robotics Society, University of Richmond
    • Establishing the university’s first Robotics Society to grow the robotics and engineering community.
    • Designing workshops on Imitation Learning (Behavior Cloning and DAgger), Model Predictive Control, and PID control.
  • 2025 - 2026

    Edinburgh, Scotland, UK

    Reinforcement Learning Team Member
    HumanEd, University of Edinburgh
    • Completed reinforcement-learning training and developed robotic-control agents with Stable-Baselines3 and PyBullet.
    • Contributed to an SO-100 robotic-arm project for real-world reinforcement-learning deployment.

Education

  • 2023 - 2027

    Richmond, VA, USA

    Bachelor of Science
    University of Richmond
    Computer Science
    • Richmond Scholar and Robins Science Scholar.
    • Dean’s List, all semesters.
    • Coursework: artificial intelligence and machine learning in robotics, algorithms, data structures, software development, probability, linear algebra, multivariate calculus, classical mechanics, and mathematical methods in physics.
  • 2025 - 2026

    Edinburgh, Scotland, UK

    Full-year Visiting Student
    University of Edinburgh
    School of Informatics
    • Coursework: Robot and Reinforcement Learning, Computer Vision, Accelerated Natural Language Processing, Control and Instrumentation Engineering, Sensor Networks and Data Analysis, Human-Computer Interaction, and Computer Security.

Publications

Projects

  • From Dynamic Programming to Reinforcement Learning
    • Compared dynamic programming, MPC, iLQR, and tabular Q-learning for stabilization, disturbance recovery, computational efficiency, and model mismatch.
    • Studied value-function discretization, MPC horizon selection, Q-learning schedules, and pole-length generalization.
  • On-Policy vs. Off-Policy Deep Reinforcement Learning
    • Compared PPO and SAC on MuJoCo locomotion and robotic-manipulation tasks.
    • Evaluated sample efficiency, stability, hyperparameter sensitivity, final policy quality, and practical compute cost.
  • Tissue Segmentation and Nuclei Classification
    • Developed U-Net-based tissue segmentation models for class-imbalanced melanoma histopathology images, comparing end-to-end training with autoencoder pre-training.
    • Built an EfficientNet-B0 nuclei classifier and evaluated ImageNet initialization, SimCLR, and supervised contrastive pre-training.
  • Cartpole Trajectory Tracking with Model Predictive Control
    • Designed an MPC controller for breathing sine-wave trajectory tracking while maintaining pole stability.
    • Completed more than 200 tuning iterations and achieved position RMSE of approximately 0.024 m in the low-frequency experiment.
  • Transformer-Based German-English Neural Machine Translation
    • Trained an encoder-decoder Transformer and analyzed generalization to longer out-of-distribution sentences.
    • Performed manual error analysis, attention-coverage analysis, BLEU and ChrF evaluation, and decoding-strategy optimization.

Honors and Awards

  • 2023
    Richmond Scholar
    University of Richmond

    One of 25 students selected from more than 11,000 applicants.

  • 2024
    Robins Science Scholar
    University of Richmond

    Merit-based award given to fewer than 3% of each class.

  • 2026
    Robert E. Loving Book Award
    University of Richmond

    Awarded to juniors with the top GPA in their class.

  • 2024, 2025
    University of Richmond Summer Fellowship
    University of Richmond
  • 2023
    Dean's List
    University of Richmond

    Awarded every semester.

Skills

Programming: Python, C++, Java, Dart
Robotics and Reinforcement Learning: ROS 2, PyBullet, CasADi, Gymnasium, Stable-Baselines3, PyTorch, Scikit-Learn, OpenCV, MediaPipe
Scientific Computing: NumPy, SciPy, Pandas, Matplotlib, Mathematica
Development and Hardware: Git, Linux, Flutter, Arduino, circuit design, sensor integration