| Research & Competition Recognition |
LAM Research Challenge 2025
Finalist
Top 0.3%
(24 out of ~8,000)
|
Selected as a Finalist for the Automated Lunar Traverse Planner project, achieving top 0.3% placement among approximately 8,000 registrations.
The project featured advanced YOLO-based crater detection, solar-shadow-aware navigation algorithms, and a three-tier hierarchical path planning system
with science-optimal waypoint selection for autonomous lunar rover missions.
|
BITS Pilani - Computer Science
B.E. Student
Aug 2024 - Present
Pilani, Rajasthan
|
Pursuing B.E. in Computer Science with focus on artificial intelligence, machine learning, and robotics research.
Currently engaged in multiple cutting-edge research projects spanning computer vision, reinforcement learning, and embedded systems.
|
| Research Initiatives & Collaborations |
IIT Indore Research
Current
Transformer-based Architecture
|
Working with Prof. Pradeep Kundu on developing Transformer-based architectures for real-time monitoring and classification of sequential
sensor data in industrial settings, with focus on designing multi-head attention modules tailored to non-stationary time-series from industrial IoT sensors.
|
POSTMAN API Lab - BITS
Current
Neuro-Symbolic AI Research
|
Contributing to research on Zero-Shot Adaptation to Non-Stationary Safety Constraints in Multi-Agent Reinforcement Learning via Neuro-Symbolic Shielding.
Developing formal symbolic safety shields and neuro-symbolic reasoning modules for dynamic constraint satisfaction.
|
CRISS Robotics - BITS
Current
Autonomous Systems
|
Serving as Software Subsystem Developer with responsibility for autonomous control systems and ROS2 architecture.
Developed autonomous airlock control system with complex FSM logic and is implementing autonomous navigation using ArUco markers and ROS2 Nav2.
|
IEEE BITS Pilani
Current
Embedded Systems
|
Designing and implementing autonomous micromouse robots for maze-solving challenges. Engineered complete embedded systems solution with Teensy 4.1 MCU,
ToF sensors, and IMU integration for real-time navigation and control.
|
Deepfake Detection
2025
87.69% Accuracy
|
Achieved 87.69% accuracy and AUC-ROC ≈ 0.90 on deepfake detection task using hybrid CNN-LSTM architecture.
Evaluated with precision, recall, and F1 scores showing robust balanced performance (F1: 88.28% for Real, 87.04% for Fake) on 2,502 video sequences.
|
Image Segmentation
2025
U-Net Architecture
|
Implemented complete semantic segmentation pipeline with U-Net achieving high Dice and IoU scores.
Integrated modern practices including GPU acceleration, reproducible seeding, Albumentations augmentation, and hybrid loss functions.
|
CIFAR-10 Classification
2024
87% Accuracy
|
Developed and fine-tuned CNN model achieving 87% validation accuracy through systematic hyperparameter tuning and transfer learning.
Implemented dropout and batch normalization for robust generalization.
|
Mechanistic Interpretability
2025
Transformer Analysis
|
Built inference and interpretability pipeline for Baby Dragon Hatchling (BDH) transformer architecture.
Implemented mechanistic interpretability probes generating high-resolution attention heatmaps to visualize model decision pathways across layers.
|
| Technical Specializations |
|
Deep Learning & Computer Vision: Advanced expertise in CNNs (ResNet, U-Net), Transformers, LSTM/GRU architectures,
neural style transfer, GANs, and vision-language models. Proficient in PyTorch and TensorFlow frameworks.
Robotics & Autonomous Systems: Hands-on experience with ROS 2, SLAM algorithms, sensor fusion, path planning,
and embedded control systems. Familiar with Gazebo simulation and real-time embedded programming.
Machine Learning Research: Mechanistic interpretability, reinforcement learning, multi-agent systems,
and neuro-symbolic AI approaches. Experience with experimental design and reproducible research practices.
Data Science & Analysis: Time-series analysis using FFT, data augmentation strategies, statistical evaluation metrics,
and comprehensive data pipeline development.
|
|