Experience & Professional Background
Current Role
Incoming MSc Machine Learning Student (Starting Winter 2025)
University of Tübingen, Germany
ELIZA Master’s Scholarship Recipient
- Selected for the prestigious ELIZA Master’s Scholarship program
- Focus on advanced machine learning, AI security, and trustworthy AI
- Research interests: Computer Vision, Medical Image AI, AI for Social Good
Previous Experience
Associate ML Scientist - 2 (January 2025 - September 2025)
Wadhwani Institute for Artificial Intelligence, Delhi, India
- Advanced Computer Vision: Developing computer vision models for crop monitoring using Joint Positional Embedding models
- MAE Pretraining: Applying Masked Autoencoder style pretraining strategies for region-specific pest recognition
- Vision Language Models: Evaluating and fine-tuning Open Source VLMs for agricultural applications
- Project Leadership: Leading pest monitoring projects with teams of ML analysts and interns
- Experimental Frameworks: Establishing frameworks for low-resource environments and noisy data challenges
Associate ML Scientist - 1 (July 2023 – January 2025)
Wadhwani Institute for Artificial Intelligence, Delhi, India
- Automated Model Development: Created standardized training and evaluation pipelines, reducing development time from 5 hours to under 1 hour
- Streamlined Deployment: Revamped deployment pipeline, accelerating releases from 5 days to under 1 day
- Enhanced Model Reliability: Improved robustness using dataset cartography techniques and increased explainability with Class Activation Maps (CAMs)
- Crop Yield Estimation: Developed and deployed an auditable system with agronomist-validated feature importance
- CottonAce Project: Integrated synthetic image generation, significantly improving model performance for national deployment within the Ministry of Agriculture and Farmers Welfare’s National Pest Surveillance System
ML Intern (February 2023 – July 2023)
Wadhwani Institute for Artificial Intelligence, Delhi, India
- Model Distillation: Developed and evaluated novel model distillation techniques for object detection and counting models
- Performance Optimization: Optimized model performance and efficiency through various distillation strategies
- Cascade Models: Analyzed and improved cascade models including OOD detectors and object detection models for edge devices
- Pipeline Development: Contributed to robust pest infestation analysis pipeline development
Research Intern (August 2022 – February 2023)
Indian Institute of Technology Delhi
- MAVI Project: Worked under Prof. Chetan Arora and Prof. M. Balakrishnan on optimizing object detection models for Mobility Assistance for Visually Impaired
- Medical AI: Worked under Prof. Chetan Arora on explaining model predictions for gall bladder cancer (GBC) detection from ultrasonography images
- Research Focus: Vision Transformers explainability, Weakly supervised object detection, and Multiple instance learning
Summer Research Fellow (June 2022 – August 2022)
Indian Institute of Technology Delhi
- MAVI Project: Optimized deep learning models for scene-text recognition using edge devices
- Model Optimization: Refactored and upgraded existing models to latest firmware versions (OpenVINO 2022, PyCam 2.0)
- VisionLAN Integration: Converted and tested Vision Transformer + Language Model to ONNX and OpenVINO IR formats
Education
Bachelor of Technology, Computer Science Engineering
Guru Gobind Singh Indraprastha University, Delhi
July 2019 – June 2023
GPA: 9.6/10
Key Achievements
- ELIZA Master’s Scholarship: Selected for prestigious scholarship at University of Tübingen
- National Impact: Contributed to Ministry of Agriculture’s National Pest Surveillance System
- Research Publications: Multiple publications in computer vision and AI applications
- Model Optimization: Reduced development time by 80% and deployment time by 80%
- Agricultural AI: Deployed AI systems used by thousands of farmers across India
Technical Expertise
Core Skills
- Machine Learning: Deep Learning, Computer Vision, Model Optimization
- Programming: Python, PyTorch, TensorFlow, OpenCV, ONNX
- Deployment: Docker, Kubernetes, Edge Computing, OpenVINO
- Research: Model Explainability, Uncertainty Quantification, Robust AI
Domain Expertise
- Agricultural AI: Crop monitoring, pest detection, yield estimation
- Medical AI: Diagnostic assistance, image analysis, explainable AI
- Assistive Technologies: Computer vision for visually impaired
- AI Security: Trustworthy AI, robust systems, uncertainty quantification
Research Focus Areas
- AI Security & Trustworthy AI: Developing robust and secure AI systems
- Computer Vision: Advanced deep learning for image analysis
- Medical Image AI: Healthcare applications and diagnostic assistance
- AI for Social Good: Agricultural technology and assistive technologies
- Model Explainability: Interpretable AI for real-world deployment