Network automation tools, Python applications, and AI/ML experiments
This write-up focuses on the engineering patterns behind a production-style simulation system: Pydantic-validated actions, fixture-driven determinism, weighted scoring, transparent reward breakdowns, and resilient inference orchestration. If you want an AI workflow that is testable, explainable, and easy to operate, these are the building blocks.
This introduces a guardrailed, AI-assisted network operations framework designed to make automation both safe and practical. It combines natural-language workflows, reusable skills, structured tool integrations, and strict read/write controls to reduce operational risk while improving speed. The result is a system that is easy to adopt, simple to extend, and production-ready for real-world NetOps teams.
Built a Malayalam-inspired programming language with a fully functional browser-based interpreter, featuring a custom lexer, AST parser, and execution engine — deployed live via GitHub Pages to make coding more culturally engaging and accessible.
Developed and published a Python-based CLI network diagnostic tool on PyPI that performs DNS resolution checks, ICMP ping latency tests, and TCP port connectivity validation — all accessible via a single command-line interface.
Developed a modular NLP API platform using FastAPI and Ollama that exposes text generation, summarization, classification, translation, Q&A, and persistent chat capabilities through structured REST endpoints with built-in Swagger documentation.
Built a fully automated tech news aggregator that fetches daily headlines from the GNews API, updates a structured JSON dataset via GitHub Actions, and dynamically displays the content through a responsive GitHub Pages deployment.
Built a fully automated “Quote of the Day” web application that fetches daily inspirational quotes from the ZenQuotes API, updates content via GitHub Actions, and displays it in a colorful, responsive interface with one-click WhatsApp sharing.
A comprehensive collection of end-to-end data science case studies covering exploratory analysis, statistical inference, SQL analytics, feature engineering, and predictive modeling using linear and logistic regression on real-world business datasets.
Built an automated weather tracking dashboard that fetches daily weather data from the OpenWeatherMap API, stores it into a version-controlled CSV, and visualizes temperature/humidity trends using interactive Chart.js charts — fully automated via GitHub Actions.
Built a multi-module AI/ML Playground featuring practical implementations of NLP, computer vision, and speech tasks—including text generation, summarization, object detection, segmentation, image captioning, VQA, ASR, and text-to-speech—structured as a reusable experimentation suite.
Built and deployed a Streamlit-based AI chatbot using Hugging Face Transformers and Meta Llama-3.2-1B-Instruct, featuring real-time conversation history, adjustable response controls (temperature/max tokens), GPU-aware execution, and secure token management via environment secrets.
Developed a real-time object detection web application using YOLOv8 and Streamlit, capable of detecting objects in images, videos, and live webcam feeds with dynamic bounding box visualization and optimized model inference.
Built an interactive real-time sentiment analysis web application using Streamlit and VADER, enabling dynamic text evaluation with sentiment scoring, classification, and a customized user interface.
Built a Python-based automation system that performs scheduled SSH configuration backups of network devices using Netmiko, with structured storage, exception handling, and failure logging for operational reliability.