Ilesh Patel

• –Led the hardware intern team in researching and selecting industrial-grade sensors and devices for STP monitoring.
• –Conducted technical analysis of water treatment methodologies including biological, chemical, and filtration processes.
• –Evaluated water quality sensors (pH, turbidity, TDS, dissolved oxygen, flow sensors) for real-time monitoring applications.
• –Studied control units and automation systems including microcontrollers and PLC-based architectures for process regulation.
• –Assisted in designing a hardware framework for integrating sensors with monitoring and data collection systems.

• –Built a document question-answering system using Retrieval-Augmented Generation (RAG) with embeddings to generate context-aware responses.
• –Implemented text preprocessing, chunking, and semantic search using vector embeddings to improve retrieval relevance.
• –Developed a Variational Autoencoder (VAE) for feature learning and anomaly detection, improving data representation quality.
• –Evaluated model performance using standard metrics such as accuracy, reconstruction loss, and precision@k.
• –Integrated trained ML models into a FastAPI backend to provide prediction APIs for web-based applications.
• –Used Docker for local deployment and version control to ensure consistent development environments.

AI-driven resume parsing and career intelligence platform providing role matching, ATS scoring, skill gap detection, and career path recommendations using Sentence-Transformers.

• –Built with Python, FastAPI, Streamlit, and Sentence-Transformers for semantic resume scoring.
• –Instant AI-driven insights on ATS score, role matching, missing skills, and career path recommendations.
• –Extracts structured text from uploaded PDF/DOCX resumes for NLP analysis.

Human-in-the-loop logistics decision intelligence system assessing pre-dispatch shipment risk using Decision Tree Classifier and real-time Weather API integration.

• –Human-in-the-loop operational control tower evaluating delivery risk before dispatch.
• –Features 3 role-based views (Seller, Manager, Supervisor) with override tracking and audit logging.
• –Combines Decision Tree ML model, WeatherAPI live data, and operational constraints.

AI-powered financial forecasting web app utilizing LSTM neural networks for deep learning stock price prediction with moving average fallback and VADER sentiment analysis.

• –Predicts stock market trends using pre-trained TensorFlow/Keras LSTM neural networks.
• –Includes dynamic fallback trend analysis using 50-day & 200-day moving averages.
• –Real-time market news integration with VADER sentiment scoring and interactive Plotly candlestick charts.