Hi, I'm Akram

Problem

Planning a trip end-to-end usually means juggling several disconnected tools and a lot of manual back-and-forth to research, compare and finalize an itinerary.

Solution

A multi-agent system built with LangGraph, where specialized agents collaborate to research and assemble a travel itinerary. Agents use tool calling to pull in external information, with a human-in-the-loop checkpoint so the user can review and adjust the plan before it's finalized.

Architecture

An orchestration graph coordinates specialized agents (e.g. research, itinerary planning) through LangGraph, with tool-calling nodes for external actions and an explicit human approval step in the loop.

Key Features

Multi-agent orchestration, LLM tool calling, human-in-the-loop approval before finalizing output, and a Python-based agent pipeline.

Tech Stack

Problem

Farmers often lack quick, data-driven guidance on which crops best suit their soil and weather conditions, leading to guesswork in crop selection.

Solution

A machine learning–based recommendation engine that takes soil and weather parameters as input and predicts suitable crops.

Key Features

Soil parameter input, weather data integration, ML-based prediction, and a simple recommendation output.

Tech Stack

Problem

Aviation data and history can feel dry and inaccessible when presented as plain tables or text.

Solution

An interactive, narrative-driven experience that visualizes aviation data to make it easier to explore and understand.

Key Features

Interactive visualizations, storytelling-driven navigation, and a fully responsive layout.

Tech Stack