Electronics & Instrumentation Engineer at NIT Silchar with a deep passion for Machine Learning, AI systems, and full-stack development. Knight on LeetCode. Builder of smart diagnostic tools.
|
Smart Disease Prediction System — TF-IDF + cosine similarity retrieval to predict top 5 probable diseases from natural language symptoms. Adaptive questioning via decision trees improved diagnostic precision by 15%. Curated ~1,151 entry dataset from Mayo Clinic.
AI-Driven Multi-Modal Art Generator using a 3-model pipeline (CLIP Interrogator → T5 → Stable Diffusion). Generates 1024×1024 images blending uploaded visuals with text prompts. Reduced inference latency by 62.5% (240s → 90s), achieving 81.7% feature retention.
Benchmarked Random Forest, XGBoost, and Neural Networks for health scoring — Random Forest achieved Test MSE = 0.0770, reducing error by 50.7% vs Neural Networks. Integrated BLOOMZ-560M chatbot fine-tuned on MedQuAD (47k Q&A pairs) for domain-specific recommendations across diet, sleep, fitness, and treatment.
Conducted ML workshops and coding sessions for 200+ students covering supervised learning, NLP, and computer vision. Mentored juniors in hands-on projects. Co-organized Neurathon 2025 — a North-East level hackathon with 100+ participants building AI/ML solutions.
Completed modules in Supervised ML, Advanced Algorithms, and Unsupervised Learning. Implemented regression, classification, and clustering models using Python, TensorFlow, and scikit-learn.
A blend of low-level systems understanding from EIE and high-level ML/AI engineering. Comfortable from embedded hardware to deploying neural networks and web applications.
Eager to contribute and grow through opportunities in Software Engineering, Data Analytics, and AI/ML.