BANOTH Siddhu

I’m an AI/ML Engineer with end-to-end experience in machine learning, NLP, and Generative AI, and I’m energized by turning models into real, usable applications. I’ve built an Agentic RAG pipeline, deployed sentiment analysis with confidence scoring, and delivered strong predictive performance for real-world-style problems.

During my AI/ML Engineer internship at Datavalley.ai in Hyderabad, I designed an Agentic RAG pipeline for restaurant review search using LangChain, ChromaDB, and Sentence Transformers. I owned data ingestion, embedding generation, and metadata filtering, then brought it to life with a Streamlit frontend that supports natural language querying.

I also fine-tuned and deployed a DistilBERT-based sentiment classifier for real-time review understanding, including confidence scoring and delivery through Hugging Face Spaces. To make the solution integration-ready, I built and exposed REST APIs using FastAPI for model inference endpoints, while training and evaluating multiple ML/DL approaches with preprocessing, feature engineering, and metrics like Accuracy, Precision, Recall, F1, and ROC-AUC.

Beyond the internship, I implemented a Loan Eligibility Prediction System using Decision Tree and Random Forest classifiers, achieving 85% prediction accuracy through a full preprocessing and feature engineering pipeline. I’m currently completing a B.Tech in Computer Science & Engineering (AI/ML specialization) and I bring hands-on learning from training and certifications, including cloud computing and ROS-based robotics, with a focus on building reliable, deployable AI systems.