Starting with a bold idea, I methodically untie knotty problems, translate them into reliable backend architecture, and present the results through joyful, lightning-fast user experiences.
Doot Agent
doot-agent-query1.mp4 |
A sophisticated agentic orchestration platform built with LangGraph for autonomous task execution. Features advanced AI agent architecture with checkpointing, interrupts, real-time streaming, memory management, and dynamic tool selection. Building specialized agents for browser automation, MCP integration, and multi-agent company workflows. Stack: Next.js, FastAPI, Python, Supabase, LangGraph, AI Orchestration, Docker microservice. |
✏️ Sketch AI
🏥 Virtual Patient Simulator
virtual_patient.mp4 |
A medical training tool that simulates realistic patient responses using a fine-tuned Llama model to help medical professionals practice patient interactions. The system uses LoRA fine-tuning to adapt pre-trained models for medical dialogue while preserving general language capabilities. Tech Stack: Python, Llama, LoRA Fine-tuning |
🛒 Smart Shopper
smart_shopper.mp4 |
An intelligent Django REST API that provides near real-time product search using SerpAPI's Google Shopping results. Optimized with async HTTP calls, Redis caching for performance, and built-in analytics tracking to monitor query patterns and cache efficiency. Tech Stack: Django, SerpAPI, Redis, HTTPX, Python |
🍷 Conversational Concierge
conversation_agent.1.mp4 |
A lightweight LangGraph conversational agent built for a Napa Valley wine business with Streamlit UI. The agent answers questions from uploaded documents, performs web searches, and provides weather forecasts. Features modular tool architecture and real-time tool call visualization. Tech Stack: Python, LangGraph, Streamlit, LLM |
React.js, Next.js, Node.js, FastAPI, Flask, Supabase, LangGraph, Gemini, Spring boot, MongoDB, Redis, PostgreSQL, Docker, AWS, Google Cloud, Microservices, Kafka, Git.

