CORAC Engineering · 2025–2026

DeepTrace.

A zero-trust data provenance platform for Earth Observation data. It verifies file lineage and integrity in the browser, without ever transferring the sensitive files.

Role: Senior Frontend Engineer
Timeline: Apr 2025 – Aug 2026
Sector: GovTech · Aerospace

DeepTrace lineage view: a flood extent product traced back through processing steps to a raw satellite downlink, with a trust mark per organization and one untrusted source flagged, demo data

The challenge

Earth Observation data is sensitive and massive. Providers and consumers need to verify that nothing was tampered with between collection, processing and delivery, but the files themselves often can’t move across networks.

DeepTrace solves this by doing the verification client-side: the browser fetches a cryptographic hash chain, reconstructs the lineage graph, and validates each step against trusted sources, without the operator ever holding the raw data.

What I owned

Lineage visualization in D3.js
Interactive force-directed graph of provenance. Each node is a processing step; edges are hash-verified transforms. Supports 1000+ nodes with zoom, pan, and query-based highlighting.
Client-side hash-chain verifier
Pure-browser cryptographic integrity check. Walks the chain from leaf to root, validates each SHA-512 hash against a trusted anchor, flags any tampered link with precise diagnostics.
Zero-trust auth with Cognito
AWS Cognito OAuth flow; access tokens never reach the server. Scoped permissions for data providers vs. consumers, with session-level hash-chain verification gating sensitive operations.
Vue 3 + Vite + Shadcn stack
Set up the project from scratch. Strict TypeScript, automated linting, pnpm workspaces, CI with preview deploys. Established patterns the rest of the team adopted.