gautam Singh

I’m a Computer Science and Engineering student at MNNIT Allahabad (B.Tech, 2022–2026), currently working as a Research Intern under Dr. Dharmender Singh Kushwaha. My focus is turning AI research into reliable systems—especially where multiple modalities must work together.

In my research internship (Jan 2026–May 2026), I engineered a transformer-based multimodal deep learning framework for context-aware hate speech detection across text and images. I benchmarked RoBERTa, DeBERTa, DistilBERT, ViT, CLIP, and EfficientNet on the MMHS150K dataset, and designed attention-based fusion to improve accuracy over unimodal baselines.

Beyond modeling, I built end-to-end preprocessing pipelines for OCR extraction, image normalization, tokenization, and multimodal feature engineering using Python and OpenCV. I also evaluated experiments using ROC-AUC, confusion matrices, precision, recall, and F1-score to guide architecture selection and tuning.

I balance research with full-stack development through projects like a QR-based Retro restaurant management system and a 2D-to-3D image converter. I’ve built REST APIs, integrated Razorpay for secure payments, and supported real-time updates—then extended into depth estimation, point-cloud/mesh generation, and interactive 3D rendering.