Resume

Basics

Name Alok Raj
Label Undergraduate Student
Email alok9360@gmail.com
Phone +91-8709414158
Url https://loki-silvres.github.io
Summary A highly motivated student with a passion for robotics and artificial intelligence.

Education

  • 2022.10 - 2026.05

    Dhanbad, India

    Bachelor of Technology
    Indian Institute of Technology (IIT ISM) Dhanbad
    Computer Science and Engineering
  • 2020.08 - 2022.06

    Begusarai, India

    12th Grade
    BR DAV Public School
    High School

Awards

Skills

Programming
C++
Python
MATLAB
Linux
Git
Simulation and Visualization
Isaac Gym
Gazebo
RViz
Frameworks and Libraries
ROS/ROS2
PyTorch
OpenCV
Matplotlib
Hardware
Nvidia Orin
Depth Camera
2D LIDAR
Dynamixel Actuators
RaspberryPi
Odroid XU4

Languages

Hindi
Native speaker
English
Fluent
Japanese
Elementary

Interests

Robotics
Computer Vision
Mobile Robots
Swarm Robotics
Legged Robots
Humanoid Robots
Control Systems
Embedded Systems
Circuit Design
Microcontrollers
Sensor Interfacing
Machine Learning
Reinforcement Learning
Foundation Models
Computer Vision

Projects

  • 2024.11 - 2024.12
    Mobile-Swarm-Navigation
    Developed a scalable robot swarm for autonomous exploration and navigation in a dynamic environment.
    • Achieved dynamic semantic environmental mapping with Instance Segmentation and Stereo Depth
    • Designed a custom semantic database management system for efficient task allocation for the swarm
    • Designed a user-friendly chatbot interface for swarm control
  • 2024.07 - 2024.10
    Autonomous Driving NXP-B3RB-buggy
    Developed an autonomous driving system for the B3RB buggy using ROS2, leveraging advanced perception and control techniques
    • Implemented LIDAR and camera for lane detection, obstacle avoidance, and traffic sign recognition
    • Trained YOLOv5s with INT8 quantization, achieving real-time inference at 7 Hz
    • Achieved a 1:42 track time with Ackermann-steering control
  • 2024.05 - 2024.07
    Embedded Toy Detection for Robots
    Advanced object detection project for robotic perception using state-of-the-art machine learning techniques
    • Trained Co-DETR model for pseudo-labelling unannotated images
    • Applied curriculum learning and knowledge distillation
    • Increased mAP50 from 34% to 42%
    • Deployed model on robot using Nvidia DeepStream and ROS2
  • 2023.08 - 2024.02
    Hologlyph Bots
    Developed a robotic system with advanced perception and control mechanisms
    • Simulated bot in Gazebo with overhead camera for arena monitoring
    • Implemented Aruco detection for pose tracking
    • Developed PID control loop with camera feedback
    • Integrated inverse kinematics for holonomic drive