Mohit Nirmal

AI/ML undergraduate · Generative AI, RAG & NLP

Ahmedabad, Gujarat, India

@mohit_nirmal

LinkedIn GitHub

0 followers

🎓 B.Tech in Artificial Intelligence & Machine Learning at Nirma University · Graduating 2028

About

AI/ML undergraduate focused on Generative AI, Retrieval-Augmented Generation (RAG), NLP, and Deep Learning. Experienced in building agentic retrieval pipelines, LLM applications, machine learning systems, and production AI APIs using Python, FastAPI, vector databases, and modern ML frameworks.

Education

2028

B.Tech in Artificial Intelligence & Machine Learning

Nirma University

Artificial Intelligence & Machine Learning · 2024 – 2028

Artificial Intelligence & Machine Learning · 2024 – 2028

Skills

<3 months

C SQL RAG Bm25 Keras Python FastAPI LangChain LangGraph JavaScript TensorFlow Embeddings Mobilenetv2 scikit-learn Transformers Hybrid Search Classification LLM evaluation Neural Networks Model Evaluation Cross Validation Transfer Learning Prompt Engineering Feature Engineering Convolutional Neural Networks

Tools / apps / platforms

<3 months

Git Linux Redis Docker Render SQLite Vercel MongoDB ChromaDB PostgreSQL

Projects

Loan Approval Prediction

Python, Scikit-learn, Pandas, Machine Learning, Streamlit

Built an end-to-end machine learning classification pipeline for loan approval prediction, covering data preprocessing, feature preparation, model training, and inference. Applied feature engineering and categorical encoding techniques to transform applicant and financial attributes into model-ready features. Trained and compared multiple classification algorithms including Logistic Regression, Naive Bayes, and K-Nearest Neighbors using Scikit-learn model evaluation workflows. Developed a Streamlit application for interactive applicant input and real-time loan approval predictions using the trained ML pipeline.
Pneumonia Detection

Python, TensorFlow, Keras, MobileNetV2, Streamlit

Developed a chest X-ray classification system using transfer learning with MobileNetV2, training custom classification layers and fine-tuning the pretrained network for pneumonia detection. Applied image preprocessing and data augmentation including rotation, zoom, horizontal flipping, and spatial transformations to improve model generalization. Evaluated the classifier using precision, recall, F1-score, and confusion matrices, achieving approximately 89% accuracy and 97% pneumonia recall with decision-threshold tuning. Built a Streamlit inference application supporting chest X-ray uploads, real-time classification, confidence scores, and pneumonia probability predictions.
ResearchMind AI

Python, FastAPI, React, LangGraph, ChromaDB, BM25

Engineered an end-to-end agentic RAG platform for multi-document research with PDF ingestion, semantic chunking, vector indexing, and source-grounded conversational retrieval. Built a hybrid retrieval pipeline combining BM25 sparse retrieval and dense vector search with Reciprocal Rank Fusion and cross-encoder reranking for relevant document retrieval. Developed a LangGraph agentic workflow for query routing, decomposition, document relevance grading, query rewriting, fallback retrieval, and grounding verification. Implemented Google OAuth, JWT-protected APIs, and per-user document isolation, supporting scoped retrieval, persistent conversations, and streaming responses.

🏆 Achievements & awards

LeetCode DSA Problems

Solved 450+ Data Structures and Algorithms problems on LeetCode and achieved Pupil rating on Codeforces.
Odoo x Parul University Hackathon Finalist

Finalist at the Odoo x Parul University Hackathon.
HackeMiNED 2026 DevX Track Presentation · 2026

Presented an AI-powered workforce intelligence platform at HackeMiNED 2026 (DevX Track).
Smart India Hackathon Participant · 2025

Participated in Smart India Hackathon 2025.

Explore the Broxer community

💬 Q&A Ask about roles, salaries and interviews — and help others. 👥 Circles Join discussions on skills, companies and industries. 📰 Feed Follow updates and insights from the community.