AI Engineer building agentic systems and automating businesses end-to-end with n8n and GoHighLevel.

Desktop HMS built for a real clinic client — offline-first, local database, PDF invoicing/reports, multi-role staff workflows.

Multi-tenant voice-agent platform letting local service businesses (clinics, salons, hotels) deploy customizable phone receptionists.

Agentic AI platform for a Dubai-based client engagement, delivered with a contracted development partner.

Conversational voice agent handling taxi-booking flows end-to-end, delivered with a contracted development partner.

Retail voice assistant combining LLM function-calling with retrieval-augmented product Q&A for a client's storefront.

Automated pipeline extracting structured data from scanned documents and receipts during a company internship engagement.

Full front-end audit deliverable — written report plus before/after screenshots — for an accounting firm client.

Paid multi-language booking demo built for a client; on hold pending the client's production budget.

AI voice and chat concierge for real-time hotel room reservations, delivered as an internship project.

Knowledge-base-driven support bot handling customer service and return-flow queries for an e-commerce client.

Agentic Vision RAG for scanned PDFs, handwritten documents, and complex layouts — 7-node LangGraph StateGraph with self-correcting retrieval.

Upload 2 documents, ask questions, and watch a LangGraph ReAct agent reason across dual ChromaDB vector stores using Mistral AI.

LangGraph-orchestrated HR engine that classifies employee requests, enforces deterministic approval policy, and integrates Google Workspace + WhatsApp.

Researcher → Writer → Reviewer agents coordinated by a LangGraph state machine router — full content pipeline from query to polished output.

Preserves AI conversations across ChatGPT, Claude, Gemini, and Copilot with intelligent context compression. Live on the Chrome Web Store.

Continuous speech-to-text and text-to-speech with Mistral AI — real-time voice conversation in a Streamlit interface.

Full-stack AI shopping assistant with native LLM function-calling, live checkout, and n8n order-automation workflows.

Enterprise-grade knowledge-retrieval assistant that scrapes and refreshes its own knowledge base for document query resolution.

Recruiting agent that parses CVs via Mistral OCR and ranks candidates against job criteria across weighted evaluation dimensions.

Local desktop playground for chatting with and testing open-source LLMs fully offline, with a cloud-API mode via Mistral AI.

Real LoRA/PEFT fine-tuning pipeline (not just inference) producing a clinical-QA specialized LLM with tracked evaluation metrics.

Interactive experimentation repository exploring array manipulation and early vector-extraction techniques for multi-modal RAG.

Human-in-the-loop content curation built on n8n: research → draft → sheet queue → human approval → scheduled Telegram post.

Multi-role (buyer/seller/admin) marketplace with split-by-vendor orders, Stripe payments, and automated test coverage.

Full-stack Learning Management System: auth, course creation, progress tracking, video, payments, and its own e2e test suite.

Task manager containerized with Docker, featuring a real CI/CD deployment pipeline.

Strictly-typed TaskFlow rewrite with generics, superseding the plain-JS version.

Seat-level ticketing platform with role-based access control.

Real-time availability checks, authentication, bookings, and reviews.

Storefront demonstrating both Context API and Redux Toolkit — with async thunks for live API data — side by side.

This site — a data-driven portfolio with per-project detail pages, dark/light themes, and optimized Web Vitals.

Full-stack mobile ride-hailing app with real-time geolocation tracking, driver matching, and fare estimation.

Library and reading-goal tracker with progress tracking, built on Expo and Appwrite.

Domain and social-handle availability checker product, kept closed-source.

All-in-one AI operating system for small businesses: internal document RAG, dashboards, and workflow automation.

Multi-tenant SaaS combining CRM pipelines with project/task management, built as a personal venture.

Early-stage platform scaffold for a self-serve enterprise AI voice receptionist product.

Offline tool for image upscaling, batch processing, and resolution management.

Four classical-ML micro-apps served via FastAPI: resume screening, churn prediction, fraud detection, and chatbot NLP.

Forecasting dashboard pairing a Python model with a Next.js UI (synthetic data, no real-brand affiliation).

Web GUI wrapper for converting PDFs, Office documents, and web pages into clean Markdown.

Utility for merging and converting between PDF, PowerPoint, and image formats.

Turtle-graphics and core-syntax exercises: Pizza Deliveries, Tip Calculator, and Treasure Island.