AI-Powered Operations System

Human-in-the-Loop Blast Cell Management Dashboard

Overview

Led the development and deployment of an AI-integrated operations system for monitoring blast cell activity across cold storage facilities, transforming manual tracking into a real-time, governance-aligned decision interface.

The solution replaced whiteboard-based workflows with a centralized dashboard that translated AI sensor data into actionable, auditable insights, enabling supervisors to make faster, more reliable decisions in high-risk, time-sensitive environments.

This work operationalized AI within real-world logistics environments by embedding human-in-the-loop decision systems that balance automation, trust, and governance.

Built to support consistent decision-making across distributed operational teams.

Problem

Blast cell operations were managed through manual whiteboard tracking, creating:

• Delays in loading and unloading workflows

• High susceptibility to human error

• Limited visibility into real-time system status

• Underutilization of existing AI sensor data

• Lack of auditability and governance over operational decisions

Although AI sensors captured temperature and capacity data, the information was not structured in a way that supported timely, trustworthy decision-making.

Approach

Human-in-the-Loop System Design

Built a human-in-the-loop decision system that integrates real-time AI signals into operational workflows while preserving human judgment, accountability, and regulatory compliance.

• Embedded AI sensor data into a centralized operational dashboard

• Enabled supervisors to validate, override, and annotate AI-driven inputs

• Prioritized safety, trust, and regulatory compliance over full automation

Governance & Auditability

Established governance-aligned workflows to ensure AI-driven decisions were traceable and defensible.

• Implemented timestamped inputs and supervisor validation checkpoints

• Designed workflows supporting audit readiness and compliance requirements

• Ensured transparency of AI recommendations within operational contexts

Research & Workflow Integration

Conducted field studies, user interviews, and usability testing to align the system with real-world operational constraints.

• Identified bottlenecks in existing workflows

• Mapped AI data outputs to actionable supervisor decisions

• Adapted system design to site-specific operational variability

Scalability & System Architecture

Designed a modular system that balanced standardization with flexibility.

• Enabled site-level configuration while maintaining governance consistency

• Leveraged existing AI sensor infrastructure to reduce integration risk

• Created a foundation for extending AI-driven decision systems across operations

Key Tradeoffs

• Automation vs. Human Oversight

• Optimization vs. Explainability

• Standardization vs. Site Flexibility

• System Complexity vs. Adoption Speed

Decisions prioritized long-term operational trust, human oversight, and governance alignment over fully automated optimization.

Outcome

• Achieved a 90% improvement in operational throughput, reducing delays and increasing response time efficiency

• Reduced errors through real-time anomaly detection and alerts

• Enabled audit-ready validation of AI-driven decisions across sites

• Increased supervisor confidence and adoption of AI-supported workflows

• Established a scalable model for AI integration in operational environments

Key Capabilities Demonstrated

• AI System Design & Human-in-the-Loop Architecture

• AI Governance & Operational Compliance

• Real-Time Decision System Design

• Cross-Functional Product Leadership

• Change Management & Workflow Adoption