ML Experiment
Why Pregnancy Features Cannot Predict Baby Gender: A Machine Learning Experiment The Age-Old Question Meets Modern AI
Read article →AI · ML · AI AGENTS · RAG
I build AI Agents, RAG systems, and LLM-powered applications that retrieve knowledge, reason over context, use tools, and turn AI capabilities into practical products.
About
I'm an AI/ML Engineer focused on building practical AI systems. My work spans AI Agents, Generative AI, RAG, LLM applications, and machine learning.
I use Python, LangGraph, LangChain, FastAPI, vector databases, and modern LLM APIs to build systems that retrieve knowledge, reason over context, use tools, and deliver useful outcomes. I enjoy taking AI ideas beyond prototypes and turning them into reliable, production-oriented applications.
Skills
How I build
My AI projects combine retrieval, agentic workflows, LLM reasoning, and external tools instead of treating the model as a standalone chatbot.
Projects
End-to-end RAG system using LangGraph to intelligently route queries across document knowledge, general knowledge, and real-time web search. Features chunking, embedding generation, query rewriting, relevance grading, and Qdrant vector search — backed by FastAPI with MongoDB chat memory.
ML pipeline for detecting fraudulent job postings using TF-IDF and XGBoost, achieving ROC-AUC of 0.92. Automated preprocessing and feature engineering; mitigated class imbalance through threshold optimization, reducing false positives by 68%.
Full-stack platform analysing resumes against job descriptions using Google Gemini AI to generate personalized interview questions, skill-gap insights, and preparation roadmaps. Features JWT-authenticated REST APIs and MongoDB-backed report management.
Forecasting pipeline on 500K+ retail transactions using RFM segmentation, EDA, customer segmentation, and feature engineering. Evaluated SARIMA and ETS models with MAE and RMSE metrics to generate actionable insights for retail demand planning.
Developed a full-stack real-time chat application using React.js, Node.js, Express.js, MongoDB, and Socket.IO, enabling instant bidirectional messaging with secure user authentication and persistent chat history. Implemented RESTful APIs and WebSocket communication to support real-time message delivery, user management, and scalable client-server interactions.
Blog
Why Pregnancy Features Cannot Predict Baby Gender: A Machine Learning Experiment The Age-Old Question Meets Modern AI
Read article →Experience
Education
B.Tech in Computer Engineering · Anand, Gujarat
Interested in AI/ML Engineering, Generative AI, AI Agent, RAG, and LLM application opportunities.