{"data":{"featured":{"edges":[{"node":{"frontmatter":{"title":"Network Security Phishing Detection Pipeline","cover":{"childImageSharp":{"gatsbyImageData":{"layout":"constrained","placeholder":{"fallback":"data:image/jpeg;base64,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"},"images":{"fallback":{"src":"/static/965c33537b7a80f38afc61e88553718d/16734/IMG_1039.jpg","srcSet":"/static/965c33537b7a80f38afc61e88553718d/46660/IMG_1039.jpg 175w,\n/static/965c33537b7a80f38afc61e88553718d/63e83/IMG_1039.jpg 350w,\n/static/965c33537b7a80f38afc61e88553718d/16734/IMG_1039.jpg 700w,\n/static/965c33537b7a80f38afc61e88553718d/53dfe/IMG_1039.jpg 1400w","sizes":"(min-width: 700px) 700px, 100vw"},"sources":[{"srcSet":"/static/965c33537b7a80f38afc61e88553718d/b3c4a/IMG_1039.avif 175w,\n/static/965c33537b7a80f38afc61e88553718d/95a66/IMG_1039.avif 350w,\n/static/965c33537b7a80f38afc61e88553718d/9e1c9/IMG_1039.avif 700w,\n/static/965c33537b7a80f38afc61e88553718d/1f4cb/IMG_1039.avif 1400w","type":"image/avif","sizes":"(min-width: 700px) 700px, 100vw"},{"srcSet":"/static/965c33537b7a80f38afc61e88553718d/a9eff/IMG_1039.webp 175w,\n/static/965c33537b7a80f38afc61e88553718d/bc7a6/IMG_1039.webp 350w,\n/static/965c33537b7a80f38afc61e88553718d/f10ea/IMG_1039.webp 700w,\n/static/965c33537b7a80f38afc61e88553718d/efa95/IMG_1039.webp 1400w","type":"image/webp","sizes":"(min-width: 700px) 700px, 100vw"}]},"width":700,"height":300}}},"tech":["Python","Scikit-learn","MLflow","Docker","AWS"],"github":"https://github.com/hari-599","external":"https://github.com/hari-599"},"html":"\n- Built an end-to-end machine learning backend system to detect phishing websites using structured website features, covering ingestion, validation, transformation, training, and prediction serving.
\n- Designed a modular pipeline with MongoDB-based data ingestion, schema and drift validation, KNN-imputer preprocessing, and multi-model training for repeatable experimentation.
\n- Integrated MLflow for experiment tracking, FastAPI for batch prediction APIs, Docker for deployment consistency, and optional AWS S3 syncing for artifact management.
\n
"}},{"node":{"frontmatter":{"title":"AI-Powered Stock Assistant","cover":{"childImageSharp":{"gatsbyImageData":{"layout":"constrained","placeholder":{"fallback":"data:image/png;base64,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"},"images":{"fallback":{"src":"/static/7c447511f09a58b35b20ee2d733cc083/2b514/octoprofile.png","srcSet":"/static/7c447511f09a58b35b20ee2d733cc083/b13f7/octoprofile.png 175w,\n/static/7c447511f09a58b35b20ee2d733cc083/771b8/octoprofile.png 350w,\n/static/7c447511f09a58b35b20ee2d733cc083/2b514/octoprofile.png 700w,\n/static/7c447511f09a58b35b20ee2d733cc083/36214/octoprofile.png 1400w","sizes":"(min-width: 700px) 700px, 100vw"},"sources":[{"srcSet":"/static/7c447511f09a58b35b20ee2d733cc083/9d1e9/octoprofile.avif 175w,\n/static/7c447511f09a58b35b20ee2d733cc083/e3526/octoprofile.avif 350w,\n/static/7c447511f09a58b35b20ee2d733cc083/5fe96/octoprofile.avif 700w,\n/static/7c447511f09a58b35b20ee2d733cc083/e1e0a/octoprofile.avif 1400w","type":"image/avif","sizes":"(min-width: 700px) 700px, 100vw"},{"srcSet":"/static/7c447511f09a58b35b20ee2d733cc083/d75a4/octoprofile.webp 175w,\n/static/7c447511f09a58b35b20ee2d733cc083/eb79c/octoprofile.webp 350w,\n/static/7c447511f09a58b35b20ee2d733cc083/c91ae/octoprofile.webp 700w,\n/static/7c447511f09a58b35b20ee2d733cc083/88077/octoprofile.webp 1400w","type":"image/webp","sizes":"(min-width: 700px) 700px, 100vw"}]},"width":700,"height":305}}},"tech":["Python","Flask","APIs","Docker"],"github":"https://github.com/hari-599/financial_agent","external":"https://financial-agent-c2io.onrender.com/"},"html":"\n- Built a decision-support application that combines real-time financial data with ML-driven query processing.
\n- Designed retrieval-based pipelines to incorporate external data into model responses and improve usefulness.
\n- Developed backend services with Flask and packaged the application for deployment with Docker.
\n
"}},{"node":{"frontmatter":{"title":"Financial Sentiment Analysis Engine","cover":{"childImageSharp":{"gatsbyImageData":{"layout":"constrained","placeholder":{"fallback":"data:image/png;base64,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"},"images":{"fallback":{"src":"/static/074b64f26a0768efe2b887f22ec48128/2b514/image.png","srcSet":"/static/074b64f26a0768efe2b887f22ec48128/b13f7/image.png 175w,\n/static/074b64f26a0768efe2b887f22ec48128/771b8/image.png 350w,\n/static/074b64f26a0768efe2b887f22ec48128/2b514/image.png 700w,\n/static/074b64f26a0768efe2b887f22ec48128/36214/image.png 1400w","sizes":"(min-width: 700px) 700px, 100vw"},"sources":[{"srcSet":"/static/074b64f26a0768efe2b887f22ec48128/9d1e9/image.avif 175w,\n/static/074b64f26a0768efe2b887f22ec48128/e3526/image.avif 350w,\n/static/074b64f26a0768efe2b887f22ec48128/5fe96/image.avif 700w,\n/static/074b64f26a0768efe2b887f22ec48128/e1e0a/image.avif 1400w","type":"image/avif","sizes":"(min-width: 700px) 700px, 100vw"},{"srcSet":"/static/074b64f26a0768efe2b887f22ec48128/d75a4/image.webp 175w,\n/static/074b64f26a0768efe2b887f22ec48128/eb79c/image.webp 350w,\n/static/074b64f26a0768efe2b887f22ec48128/c91ae/image.webp 700w,\n/static/074b64f26a0768efe2b887f22ec48128/88077/image.webp 1400w","type":"image/webp","sizes":"(min-width: 700px) 700px, 100vw"}]},"width":700,"height":305}}},"tech":["Python","PyTorch","Hugging Face"],"github":"https://github.com/hari-599/Stock-predictor","external":null},"html":"\n- Developed NLP pipelines for financial sentiment classification using transformer-based models including BERT and RoBERTa.
\n- Designed experiments to compare model performance against baselines and improve classification accuracy through iterative evaluation.
\n- Built modular preprocessing, training, and evaluation workflows to support repeatable experimentation.
\n
"}}]}}}