The problem Retr.ai solves

A Broken Hiring System in Indonesia

Indonesia's hiring ecosystem is fraught with **inefficiency, bias, **and a profound lack of trust. The consequences are staggering.

For Companies

A significant talent skills gap persists, with over 30% of businesses struggling to find qualified talent despite a vast workforce [1]. This leads to longer time-to-hire, high turnover, and the substantial financial costs of a bad hire. The traditional hiring process is also prone to human bias, making it difficult to objectively evaluate candidates.

For Job Seekers

Candidates are often left in the dark about their performance, receiving little to no feedback. They lack a credible way to showcase their validated skills and interview readiness, making it difficult to stand out in a competitive market. Furthermore, a growing concern over data privacy in Indonesia, underscored by the new Personal Data Protection (PDP) Law (UU No. 27 of 2022), creates a fear that their sensitive personal data could be misused or compromised on centralized platforms [2, 5].

Our Web3 Solution: Retr.ai

Retr.ai is a dual-sided, AI-powered platform that leverages Lisk to fix this. It provides a secure, trustless system that empowers both sides of the hiring market.

For Companies

Retr.ai offers AI-driven interview simulations that provide deep, unbiased, and data-backed insights into an applicant's skills. This saves time, standardizes the evaluation process, and allows companies to focus on quality candidates. Most importantly, it mitigates legal and reputational risks associated with the new PDP Law by providing a verifiable, compliant method for accessing candidate data with explicit consent.

For Job Seekers

Candidates use our personal AI coach to get detailed, actionable feedback on their performance. The highlight is a self-sovereign, verifiable career passport. This is not a simple PDF; it's an immutable, blockchain-backed record of their validated skills and interview performance, owned entirely by them. They can grant permission to companies to view this data, giving them control and a powerful, verifiable credential to showcase their readiness.

By making AI insights verifiable and giving candidates true ownership of their data on a Lisk sidechain, Retr.ai turns a low-trust system into a transparent, efficient, and equitable one. We make Web3 for everyone by seamlessly integrating it into a familiar application, solving a real-world problem for at least 19 millions of people in Indonesi just like what our vice president promised.

Challenges I ran into

Structuring and Storing Verifiable Credentials on-chain

My initial idea was to store all of the AI-generated interview feedback directly on the Lisk sidechain to make it fully (maybe, naively) decentralized. This quickly became impractical when we cale it up just a bit. Not only would it be expensive to store large amounts of text data on-chain, but it would also be inefficient and violate a core principle of good blockchain architecture.

I adopted a more sophisticated and scalable approach. Instead of storing the full feedback report on-chain, I now use the Lisk sidechain to store a cryptographic hash of the report. The full, encrypted report is stored on a decentralized storage network (IPFS. Technically, it's blockchain-like). The hash on the blockchain serves as a tamper-proof and verifiable pointer to the data. This allows the candidate to prove the authenticity of their report to any third party (like a company) without exposing the entire report's contents on the public ledger, thus ensuring both privacy and verifiability at a low cost.

Xellar Kit Integration to the Available Project

Somehow, this problem returned. I read the docs, followed the basic example, and it still didn't work. Currently still debugging this issue. Something that should have been simple, turned into a bunch of resultless debugging.

Not got over yet.