Skip to content
Wildan.
Back to work

Public Beta

Fradium

Making on-chain risk legible before money moves.

Role
Team Lead · Full-Stack Developer
Year
2025-2026
State
online

Fradium began with a simple question: what evidence should a user see before sending funds to an unfamiliar address? I led the team and worked across the full stack to turn that question into a public product spanning address analysis, canister-hosted AI, community signals, wallet touchpoints, and an explicit user decision.

Fradium product cover showing its Web3 transaction trust layer, wallet analysis surfaces, and public website.

The problem

Crypto users often reach transaction confirmation with fragmented context about a destination address. The most consequential decision can happen before the available evidence has become legible.

Built for

Crypto users evaluating an unfamiliar destination address

Community reporters and validators contributing address evidence

Wallet and dApp developers exploring analysis integrations

How it works

  1. 01

    Detect the chain and prepare chain-specific evidence.

  2. 02

    Run the embedded ONNX model through an Internet Computer canister.

  3. 03

    Keep model output and community evidence conceptually distinct.

  4. 04

    Render the verdict before the user cancels or explicitly confirms the action.

What I built

Led product direction, team coordination, and multidisciplinary delivery.

Worked across the web application, wallet and extension touchpoints, and canister-backed flows.

Connected risk analysis to the transaction decision instead of leaving it in a separate dashboard.

Shaped the public-beta demonstration and technical narrative.

Fradium was built by a multidisciplinary team. The WCHL result belongs to the Fradium team; this case study foregrounds the product direction and full-stack responsibilities Wildan personally led.

Key decisions

Put the risk signal inside the decision

Evidence becomes actionable when it appears at the transaction boundary.

Tradeoff: The wallet flow needs more explanation without taking final control away from the user.

Keep model and community signals distinct

A single opaque score would hide disagreement between evidence sources.

Tradeoff: Separate signals require clearer moderation and interface language.

Results & sources

TEAM RESULTAS OF 2025

The Fradium team won the WCHL 2025 Global Finale Fully On-Chain Track.

Telkom University
PROJECT RECORDAS OF 2026-07-15

Fradium is publicly accessible at fradium.io.

Live public beta
PROJECT RECORDAS OF 2026-07-15

The reviewed implementation combines chain preparation, canister inference, and community cross-checking.

Canonical repository