Research Intern (Deepfake Detection) at Indian Institute of Management Ranchi (2025-06 – 2025-09)
Guide: Prof. Sobhan Sarkar
- Proposed a lightweight dual-stream neural architecture for robust deepfake detection under real-world distortions
- Evaluated robustness across multi-level augmentations with low computational overhead for edge deployment
- Incorporated Grad-CAM based interpretability, analyzing model attention shifts under varying intensities
Research Intern (Analysis of Agile Teaching) at Victoria University of Wellington (2025-09 – 2025-12)
Guide: Prof. Jennifer Ferreira
- Processed 2k+ Agile teaching excerpts using TF-IDF, Latent Dirichlet Allocation and K-Means clustering
- Extracted 10 topics from positive and negative corpora and applied PCA and t-SNE for pattern visualization
- Incorporated findings from 70+ research papers to validate multiple themes and conduct temporal trend analysis
- Drafted a research paper summarizing methodologies, experiments and key conclusions to contribute to academia
Data Science Intern at Finosauras (2025-05 – 2025-06)
Received a Letter of Recommendation (LoR) from Director, highlighting exceptional professionalism and dedication
- Engineered Django dashboard analyzing 500+ articles daily extracting stock tickers and insights with 90% accuracy
- Developed end-to-end NLP pipeline after evaluating 7 LLM APIs and using A/B testing, reducing manual effort
- Integrated MongoDB and manual validation, boosting data reliability by 70% and enabling real-time trend analysis
Web Development Intern at Global Sustainability Solutions (2025-06 – 2025-08)
- Developing 3 ESG platforms, 10+ dashboards aligned with 5+ frameworks; improving reporting efficiency by 90%
- Building 15+ RESTful APIs for workflows, and ESG data; improving backend speed by 30% via query optimization
- Prototyping 10+ AI/ML pipelines for text classification, risk extraction from 100+ reports and time-series forecasting
Generative AI Research Intern at MapIT.ai (2024-12 – 2025-01)
- Custom-trained domain-specific language model for improving food-ordering accuracy by 50%, boosting purchases
- Developed FastAPI app with fine-tuned models, Dockerized for scalability, enhanced capability for 10+ languages
- Applied few-shot prompting, hyperparameter tuning and RAG techniques to boost multi-agent system performance
Junior Software Engineer at Team Rakshak | Institute Technical Council (2024-02 – 2025-01)
Advisor: Prof. Krishnendu Haldar | Team of 50+ students that develops UAVs to support Search and Rescue Operations
- Engineered Super-resolution techniques, enhancing image quality for object detection pipeline in aerial drones
- Implemented the Soft Actor-Critic(SAC) algorithm in Reinforcement Learning for autonomous navigation
- Developed proficiency in Neural Networks and advanced Image Processing using CNNs, Pytorch and OpenCV
