Election Integrity Platform

Overview

Daniel led the engineering of an end-to-end election-integrity platform that digitizes the full election workflow, from voter and party-agent accreditation to national result collation. Its defining feature: handwritten results captured at polling units are read by machine and automatically reconciled up the chain, replacing slow, error-prone manual tallying with an auditable, real-time pipeline.

The problem

Elections in the region are still tallied from handwritten result sheets that move by hand from polling units to wards, states and finally the national level. Every manual transcription and re-entry is a point where numbers can drift, get transposed, or be disputed, and the delay between voting and a credible result erodes public trust. The challenge was to preserve the paper trail officials rely on while making the count fast, verifiable, and hard to tamper with.

Approach

1. Modelled the real election workflow end to end

   Built the full lifecycle in software — voter and party-agent accreditation, result entry, and multi-stakeholder reporting — so the platform mirrored how a real election actually runs rather than forcing officials into an unfamiliar process.

2. Applied handwritten character recognition to the source documents

   Integrated HCR/OCR to extract figures directly from handwritten polling-unit result sheets, keeping the original paper record as the source of truth while removing the manual re-typing that introduces transcription errors.

3. Built automated multi-level collation with reconciliation

   Designed the aggregation engine that rolls results up through every tier — polling unit to ward to state to national — reconciling figures at each level so discrepancies surface immediately instead of being discovered days later.

4. Delivered real-time reporting for stakeholders

   Streamed collated results into live reporting views so decision-makers could watch the count progress in real time, with the accreditation and collation data feeding a single consistent picture.

5. Engineered for reliability under pressure

   Architected the system on a production-grade Next.js and PostgreSQL stack with a Node backend, prioritizing data integrity and auditability so results held up to scrutiny — in line with Daniel's principle of building systems that don't fail.

Impact

• Replaced manual, multi-hop transcription of results with an automated pipeline that reads handwritten sheets directly, cutting a major source of human error.
• Delivered real-time visibility into the count across all levels, compressing the wait between voting and a credible tally.
• Preserved the original paper result sheets as an auditable source of truth, strengthening the platform's defensibility against disputes.
• Provided reconciliation at every collation tier, so mismatches are caught the moment they appear rather than after the fact.
• Led the initiative as the responsible engineer from architecture through deployment.