Back to projects

IBM Hackathon · 2025

PulseArchitect

An autonomous enterprise AI agent that orchestrates workflow automation and measures its resource footprint.

Source code

Overview

An autonomous, enterprise-grade AI agent built on the BeeAI Framework that orchestrates and executes complex workflow automation. I developed a custom telemetry and resource-tracking layer that dynamically calculates the computational and environmental footprint — specifically water-consumption metrics and token efficiency — per incoming request sequence.

Problem

Agentic systems are accepted as long as they 'work', yet the real cognitive and environmental cost of each request — tokens, energy, water — stays invisible.

Approach

I built multi-step agent orchestration with BeeAI and added a resource-tracking layer that measures token usage and estimated water/energy footprint per request, wired into a live operator dashboard for decision-making.