Building production-grade AI systems — from LLM agents and RAG pipelines to full-stack apps. Specialised in Groq, LangChain, Gemini, Mistral and agentic architectures that actually ship.
I'm Charan Redddy Bonamukkala, a Software Engineer and AI Researcher from Hyderabad building intelligent systems at the intersection of LLMs, agentic workflows, and full-stack architecture.
I work across the full AI stack — from crafting RAG pipelines with FAISS vector stores and building multi-tool LangChain agents to deploying Streamlit apps with real-time Groq inference. I believe in shipping real things, not demos.
Currently focused on agentic AI, predictive ML models, and building tools that help people make smarter decisions faster. Follow my builds on @trending.tech.ai.
A production-grade AI agent powered by Groq's Llama 3.3 70B with live web search via Tavily. Uses a LangChain agent framework to decide when to search the web, retrieve real-time results, and synthesize structured, accurate answers. Dual interface — CLI and a clean Streamlit web app.
AI-powered trust and verification app built in Google AI Studio using the Gemini API. Detects anomalies and verifies digital identities with real-time scoring. Deployed on Google AI Studio and runnable locally with Node.js.
AI resume parser powered by Mistral 7B via OpenRouter. Extracts PDF text using pdfplumber, then uses the LLM to score candidates against job descriptions with structured, actionable insights in a Streamlit UI.
Machine learning model that predicts IPL match outcomes with 88%+ accuracy. Trained on historical IPL data using XGBoost and scikit-learn, with features like venue, toss result, team form, and run rate. Interactive Streamlit interface for live match predictions.
Retrieval-Augmented Generation system that lets users query their own documents. Chunks PDFs, embeds them into a FAISS vector store, then retrieves context for LLM answers — no hallucination, grounded responses only. Built with LangChain and Streamlit.
Full-stack TypeScript portfolio built with Vite, React, and Tailwind CSS. Production-deployed on Vercel at charan-lilac.vercel.app with clean component architecture and responsive design.
Real-time weather app with location-based forecasting, dynamic UI updates, and live API integration. Clean, responsive design with smooth data visualizations and geolocation support.
Immersive travel guide with rich media galleries, interactive destination discovery, and location-based content. Performance-first build with pure HTML, CSS, and JavaScript.