Hironmoy Roy Rudra

Undergraduate Student | AI/ML | Robotics | NLP

Dhaka, Bangladesh hironmoy5458@gmail.com +880-1611582033 rudra-roy.github.io hironmoyroyrudra Rudra-roy

Summary

Driven Undergraduate Student Researcher specializing in autonomous systems and machine learning. Proven track record of architecting robust autonomy stacks, developing high-fidelity digital twins in Unity and Unreal, and designing complex robotics pipelines for international Mars rover competitions. Currently researching parameter optimization strategies for LLMs. Highly proficient in Python, C++, ROS2, and core AI frameworks, blending rigorous academic research with hands-on robotic systems engineering.

Education

BRAC University

BS in Computer Science, Dhaka, Bangladesh

SEP 2022 – Present

  • CGPA: 3.65/4.0
  • Research: Robotics, Autonomous System, High fidelity Simulations, IOT
  • Honors: VC's List for Academic Excellence (4 Semesters), Dean's List for Academic Excellence (3 Semester)

National College of Education

Higher Secondary Certificate, Science, Narsingidi, Bangladesh

2019 – 2021

Research Experience

BRACU Mongol Tori (University Rover Team)

Self-employed / University Project Team

Dhaka, Bangladesh

Nov 2025 – Present

  • Co-Advisor (Nov 2025 – Present)
    • Direct high-level technical strategies and oversee the competitive development roadmap for BRACU Mongol Tori (ERC team), guiding the multidisciplinary engineering team through rigorous preparation for European rover challenge competition.
    • Collaborate hands-on with team members to advance applied robotics research, driving the implementation of complex autonomous navigation pipelines, robust computer vision models, and advanced ROS architectures.
    • Facilitate comprehensive knowledge transfer and mentor undergraduate engineers in advanced system design, ensuring sustainable proficiency in simulation-to-real validation and intelligent systems development.
    • Elevate the team's academic and engineering rigor by instituting strict research protocols, guiding members to parallel their competition deliverables with the high-impact standards required for top-tier, peer-reviewed robotics publications.
  • Sub-Team Lead (Sep 2024 – Present)
    • Led the Autonomous Sub-Team in developing and deploying robust autonomy stacks, SLAM pipelines, and path-planning algorithms for the Mars rover.
    • Set technical milestones, managed software development sprints, and mentored junior members in ROS2, Python, and intelligent systems.
    • Spearheaded the autonomous navigation missions for the team during international competition, University Rover Challenge (USA)
    • Managed the autonomous systems pipeline and monitored real-time telemetry during high-pressure competition tasks.
  • Control Systems Engineer (Feb 2024 – Aug 2024)
    • Designed and built ROS2-based communication and control system architecture from scratch.
    • Implemented inverse kinematics using Aruco markers and OpenCV pipeline for a 3 DOF arm.
    • Worked on efficient camera calibration techniques
  • Contributor (Nov 2023 – Feb 2024)
    • Researched and learned rover control systems (Arduino, Raspberry Pi).
    • Studied the feasibility of implementing ROS for the Mars rover system.

Participant & 6th Place Finalist (National Round)

KIBO Robot Programming Challenge (6th Edition)

Dhaka, Bangladesh

Aug, 2025 – Feb, 2026

Project: Computer Vision and Autonomous Sign Detection

  • Achieved 6th position nationwide by developing a robust computer vision pipeline for the Kibo-RPC competition.
  • Engineered algorithms for the real-time detection and classification of various types of visual signs and markers.
  • Optimized image processing models to ensure high accuracy and performance within simulated microgravity environments.
  • Integrated computer vision modules with the robotic navigation stack to execute autonomous targeting tasks.

Experience

Technical Mentor

First Global Challenge (FGC) Team Bangladesh

Dhaka, Bangladesh

Jan 2025 – Jun 2025

  • Mentored the national robotics team over a 6-month period, delivering comprehensive training in programming fundamentals, strategic planning, and logic building.
  • Guided team members in developing problem-solving strategies and robust software architectures for competition tasks.
  • Designed and implemented an end-to-end computer vision pipeline for rice leaf disease detection.
  • Leveraged foundational PyTorch models for the detection pipeline and optimized inference performance using TensorRT for high-speed edge deployment.

AI Engineer

ROBO hobby

Dhaka, Bangladesh

Oct 2024 – Jan 2025

  • Worked on edge computing for Pest control in Agricultural Settings
  • Built an AI-powered pest detection system using Raspberry Pi and TensorFlow Lite, achieving high accuracy in identifying common crop pests.

Student Instructor

Robotics Club of BRAC University

Dhaka, Bangladesh

Oct 2025 – Nov 2025

  • Conducted classes on fundamental and intermediate applications of Robotic Operating System (ROS) and fundamental physics simulations using Gazebo and Webots.
  • Provided assistance to learners by responding to queries, debugging code, and offering guidance on project development.

Projects

Hybrid RL-IL Urban Navigation Framework

Sep 2025 - Present

  • Developing a hybrid Reinforcement Learning and Imitation Learning (RL-IL) navigation framework within the CARLA simulator to tackle highly unstructured traffic scenarios.
  • Designing robust navigation and path-planning pipelines that leverage a comprehensive simulated sensor suite to ensure real-time spatial awareness and obstacle avoidance.
  • Implementing and rigorously tuning Model Predictive Control (MPC) algorithms to generate smooth, kinematically feasible trajectories and optimize vehicle control in dynamic environments.
  • Curating the Dhaka-Dense Dataset (D3) using the Segment Anything Model (SAM) for pixel-level segmentation of local vehicle classes, featuring contextual LLM-based annotations.

Unity-Based Digital Twin & Autonomous Traversal Simulation

rudra/Rover_sim

Dec 2025 - Present

  • Developed a high-fidelity digital twin of the "BRACU Mongol-Tori" Mars rover in Unity Game Engine to simulate and validate autonomous navigation pipelines.
  • Integrated ROS2 with MuJoCo physics engine, URDF joints, and custom mesh collisions to realistically model the rover's rocker-bogie suspension system.
  • Simulated rough-surface planetary environments to test path-planning algorithms and autonomous traversal capabilities without risking physical hardware.
  • Replicated complex sensor suites within the simulation to generate synthetic data for testing SLAM and LiDAR-GNSS fusion algorithms.

EchoMind: VLM Agent Wrapper

Rudra-roy/EchoMind--You-speak-It-thinks

Sep 2025

  • Developed "EchoMind," a TypeScript-based Vision-Language Model (VLM) agent wrapper designed to streamline interactions with locally hosted AI models.
  • Implemented an API-driven architecture to efficiently route user prompts to a local server and dynamically render the model's responses.
  • Managed asynchronous client-server requests to ensure low-latency communication and a seamless "speak and think" interaction loop.
  • Built a responsive interface for testing and interacting with VLM capabilities, enhancing the experimentation pipeline for local AI deployments.

Custom ROS2 Autonomous Robot Framework (my_bot)

Rudra-roy/my_bot

Sep 2024

  • Developed a custom autonomous mobile robot workspace using ROS2, featuring comprehensive URDF modeling and kinematics configuration.
  • Implemented the Nav2 stack to facilitate dynamic path planning, obstacle avoidance, and precise waypoint navigation.
  • Integrated simulated sensor data streams, including LiDAR and IMU, to enable robust Simultaneous Localization and Mapping (SLAM).
  • Configured scalable launch files to seamlessly deploy and test the robot's control algorithms within simulation environments.

ROS2 Autonomy Workspace & Unity Integration

bracu-mongoltori/simple_robot_ws

Dec 2025

  • Architected the foundational ROS2 workspace (`simple_robot_ws`) serving as the core autonomy backend for the BRACU Mongol-Tori Mars rover.
  • Successfully integrated the ROS2 control pipeline directly with a high-fidelity Unity digital twin, establishing a seamless software-in-the-loop (SITL) simulation environment.
  • Developed modular ROS2 nodes for path planning and kinematics that drive the rover's behavior within the simulated rough-terrain environment.
  • Accelerated the testing phase for autonomous traversal algorithms by validating edge-case scenarios virtually prior to physical hardware deployment.

3D LiDAR Map-Based Topographical Path Planner

Custom Navigation Algorithm

Aug 2025

  • Engineered a custom path-planning algorithm founded on Artificial Potential Fields to execute safe, autonomous waypoint traversal in complex environments.
  • Processed and analyzed raw LiDAR point cloud data to dynamically calculate terrain slopes and isolate high-risk danger zones.
  • Demonstrated superior traversal efficiency and route optimization compared to standard search algorithms (BFS, DFS, Dijkstra, Hill Climbing) and conditional A* by leveraging continuous spatial heuristics.
  • Designed the system specifically for robust mobile robot navigation, ensuring fluid obstacle avoidance and safe gradient negotiation on rough surfaces.

Skills

Programming & Databases

Python (Expert), C/C++, C#, JavaScript, TypeScript, SQL, Bash

Deep Learning & Computer Vision

PyTorch (Expert), TensorFlow, TensorRT, Keras, OpenCV, YOLO (v8/v11/world/v26), OCR (Tesseract)

AI / ML / NLP

Scikit-learn, RAG Systems, Vector Databases (Pinecone, ChromaDB), LLMs (GPT, Llama, DeepSeek), Hugging Face Transformers, Reinforcement Learning (DQN, PPO), Optuna Optimization, LoRa

Robotics

ROS2, Physics Simulation (Gazebo, Webots, PyBullet), High-fidelity Navigation simulation (CARLA, AirSim, Unity, Unreal) Path Planning, Inverse Kinematics, Control Systems Design, Sensor Integration

Embedded Systems

Arduino, Raspberry Pi, ESP32, STM32

Data Science & Analysis

Pandas, NumPy, Bioinformatics Pipelines, Data Visualization, Feature Engineering, Data Augmentation

3D Design, Simulation & Modeling

3D Design (Fusion 360, OnShape), Blender (Basic)

Web & Backend Development

Flask, FastAPI, React-Native, Tailwind CSS, Laravel, MySQL, MongoDB, Expo-Go

MLOps & Cloud

Docker, Git/GitHub, CI/CD, Weights & Biases, Model Quantization

Software Engineering & Tools

Reverse Engineering, GUI Development (PyQt, Tkinter), Linux (Ubuntu, Raspbian), Jupyter Notebooks

Professional Skills

Team Leadership, Project Management, Agile Methodologies, Critical Thinking, Technical Documentation

Language Skills

  • English: Professional proficiency
  • Bangla: Native or bilingual proficiency

Certificates & Training

Computer Science Fundamentals

Issued by: Harvard University, CS50 (2023)

100 days of code with Python

Issued by: Udemy (2023)

DSA with Python3

Issued by: Udemy (2023)

Ethical Hacking with zSecurity

Issued by: Udemy (2023)

Deep Learning Specialization

Issued by: Deeplearning.AI, Coursera (2024)

Kibo RPC 6th Edition Participant (2025)

Issued by: JAXA (Dec 2025)

Duke of Edinburge Leadership Program, Bronze Award (2023)

Issued by: The Duke of Edinburgh’s International Award Foundation’s office

Honors & Awards

8th Place, University Rover Challenge 2025

Issued by: The Mars Society, Utah, USA (Jun 2025)

21th Place, University Rover Challenge 2024

Issued by: The Mars Society, Utah, USA (Jun 2024)

National Round Participant, KIBO Robot Programming Challenge 2025

Issued by: Japan Aerospace Exploration Agency (JAXA), Tokyo, Japan (Feb 2026)

References

Dr. Md. Khalilur Rhaman, PhD

Professor, Department of Computer Science and Engineering
BRAC University, Dhaka, Bangladesh

Email: khalilur@bracu.ac.bd

Phone: +8801752042223

Dr. Farig Yousuf Sadeque, PhD

Assistant Professor, Department of Computer Science and Engineering
BRAC University, Dhaka, Bangladesh

Email: farig.sadeque@bracu.ac.bd