Systems Built for Impact

Challenge

Country-owned disaster loss and damage records are fragmented across legacy DesInventar instances and inconsistent hazard classifications; translating that record into evidence countries can act on, and into impact thresholds anticipatory action partners can trigger against, needs methodology before it needs software.

Solution

Shaping product definition and the national, regional and global instance model, the Analysis page and the Geoportal. Developed the methodology linking hazardous events, disaster events and disaster records against the Hazard Information Profiles 2025, including a cascade chain of 79 hazards and 1,193 directed links. Framing how DELTA is enriched with exposure, vulnerability, population and climate data from free and open sources, while keeping every external value outside official national totals. Contributing to the disaster data ecosystem assessment methodologies and the technical readiness assessments that review how countries produce, manage and share disaster and climate data as part of the DELTA roll-out.

Impact

• ▸85 countries in the roll-out scope
• ▸Cascade chain: 79 hazards, 1,193 directed links
• ▸HIPs 2025 methodology linking hazardous events, disaster events and records
• ▸Open-data anticipatory action pilot (Mozambique)

Challenge

Afghanistan's humanitarian response involved 200+ organizations reporting to multiple clusters with inconsistent definitions, fragmented Excel-based workflows, and no unified view of response progress.

Solution

Led the development and deployment of a complete reporting ecosystem: harmonized data standards across clusters, managed automated data collection and validation pipelines, designed role-based dashboards for different user types (cluster coordinators, OCHA, donors), and established feedback loops that improved data quality over time.

Impact

• ▸200+ organizations supported
• ▸Multi-cluster harmonization
• ▸Real-time HRP monitoring
• ▸Replaced fragmented Excel reporting

Challenge

Humanitarian actors in Afghanistan lacked a centralized geospatial platform that combined multiple hazard layers with population data, accessibility information, and response capacity.

Solution

Led the development and deployment of an integrated geospatial platform combining satellite imagery, climate data, seismic sensors, and ground-verified information. Managed the development of exposure estimation models, automated alert systems, and web-based mapping interfaces accessible to non-GIS specialists.

Impact

• ▸First integrated multi-hazard platform for Afghanistan
• ▸Automated exposure estimation
• ▸Climate trigger development
• ▸Served as national geospatial reference

Challenge

Five humanitarian organizations were conducting post-distribution monitoring independently, each with different tools, questions, and definitions of "success." The Ethiopia Cash Working Group couldn't answer: "Is our cash programming working?"

Solution

Designed a 9-pillar analytical framework covering expenditure patterns, food security impact, market access, protection, coping strategies, and beneficiary satisfaction. Built Python harmonization pipelines and a Power BI ecosystem with 50+ DAX measures.

Impact

• ▸1,559 households analyzed
• ▸5 organizations unified
• ▸9 analytical pillars
• ▸First inter-agency cash impact assessment

Challenge

IOM Bangladesh needed a structured IM system to support process-based CBI delivery ensuring timeliness, accountability, and fraud prevention across Cash-for-Work and MPCG programmes in the Cox's Bazar refugee response.

Solution

Designed the CBIIMS, a 5-stage workflow (Profiling → Targeting → Attendance → Payment Automation → Beneficiary Payment) with governance layers for risk assessment and M&E/AAP.

Impact

• ▸5-stage CBI workflow designed
• ▸Automated payment sheet preparation
• ▸Fraud prevention systems
• ▸3-step authorization process

Challenge

Humanitarian situation analysis is slow, labor-intensive, and often outdated by the time it reaches decision-makers.

Solution

Designing an agentic AI architecture where specialized agents handle data ingestion, classification, cross-source synthesis, and automated report generation, producing near-real-time analytical outputs.

Impact

• ▸AI agent architecture designed
• ▸Multi-framework classification
• ▸Real-time synthesis capability
• ▸Humanitarian-specific NLP pipeline

Challenge

Despite advances in climate forecasting, most humanitarian response in Afghanistan remained reactive. Early warnings weren't linked to pre-agreed response protocols.

Solution

Developed climate trigger thresholds calibrated to Afghanistan's hazard profile, integrating drought indices, flood forecasting, and seismic data. Created decision frameworks linking forecast levels to specific response actions.

Impact

• ▸Climate trigger thresholds developed
• ▸Early warning integration
• ▸Anticipatory action protocols
• ▸Multi-hazard risk framework

Challenge

COVID-19's impact on Bangladesh's vulnerable populations was rapidly evolving, with information scattered across hundreds of sources in multiple languages and formats.

Solution

Led a team deploying the DEEP platform to systematically code and classify thousands of unstructured humanitarian documents. Produced seven thematic reports.

Impact

• ▸7 analytical reports produced
• ▸AI-enabled data classification
• ▸Cross-sectoral synthesis
• ▸Used by humanitarian country team

Challenge

No systematic assessment existed of IM capacity gaps across humanitarian clusters, making it impossible to target capacity building efforts effectively.

Solution

Designed and deployed a multi-cluster IM capacity assessment using structured HDL surveys, producing Power BI dashboards with cluster-specific gap analysis and training recommendations.

Impact

• ▸10 clusters assessed
• ▸Capacity gaps quantified
• ▸Training priorities identified
• ▸Power BI analytical dashboard

Challenge

Contingency planning for rainy season flooding required spatial risk analysis combining hazard, exposure, and vulnerability data that did not exist in a unified format.

Solution

Created multi-layer vulnerability maps integrating flood hazard zones, settlement data, displacement sites, and infrastructure locations. Delivered as both interactive web maps and high-resolution print products.

Impact

• ▸Risk zones mapped
• ▸Contingency planning supported
• ▸Multi-hazard overlay analysis
• ▸Print and digital products

Challenge

The CWG needed visibility into CERF-funded cash programme progress across multiple partners but had no standardized tracking mechanism.

Solution

Designed Power BI dashboards with automated data pipelines tracking who, what, where, when, and for whom across all CERF-funded MPC activities.

Impact

• ▸$3.20M allocations tracked
• ▸185K beneficiaries monitored
• ▸Quarterly snapshots produced
• ▸Partner performance visibility