§ ABOUT

The Founder.

Verdict was built by a private candidate with zero resources - fed up with a market of edtech LLMs that hallucinate and hold back. Dogfooding the product development with his own exam prep and metacognition into what actually compounds tutoring results.

§ 01 - How this got built

Verdict came out of needing it.

My name is Lucas. I'll spare the lofty-sounding history of where I was enrolled (I wouldn't even say studied, because frankly I never did). Mine was an arduous journey, from holding the schools' record for conduct slips, to getting kicked out of my home for skipping a major paper - to somehow graduating JC despite sleeping through the majority of papers.

I never subscribed to the idea that academics was a path I had to take, but I developed an interest in some major and some minor university programs after I had already graduated.

The problem was the lack of support and resources. Naturally, I had zero foundational grounding for any of my H2 subjects because I spent my time in class eating chicken wings at the back. I scoured the net for resources, cross-referencing them back to the official SEAB documents, and even tried looking for LLM wrappers to guide me through the curriculum.

I was about to pay $50 out of pocket for Tutorly when I tested its free trial and realized that the syllabus alignment was completely off. It provided B-grade answers for economics at best, and when asked what were the 2026 set texts for Literature, it gave me an expired text and an O level one. I did some further digging and realized that most were either GPT/Gemini wrappers (such as Astra) or glorified information synthesizers (e.g. Penseum).

The same way we at VOSE believe education should be open source, Verdict was never designed to be constrained.

Many might've said, or even mocked, me that I wasn't much of a scholar - but I was teaching myself countless virtuosities outside of a rigid curriculum. Verdict, in similar nature, can be retrofitted for ANY educational curriculum with VOSE API+ID. An ABRSM theory guide, online CFA course prep agent, white-labelled international school LLM tutors, our wireframe is both infrastructure and philosophy.

Despite my personal historical lack of academic engagement, I've done extensive chess tutoring from children 6yrs to acquaintances older than myself. My close friends and alumni from RJC also have extensive academic tutoring experience that I've siphoned from to create an understanding of how to trigger a student's metacognitive prowess. The reverse-loading methodology, the logic gates framework, every feature isn't just nice to have - they're designed to align with how you learn.

§ 02 - What Verdict represents

Open, not closed.

I came into the production of this idea knowing full well that the edtech market was polluted with tons of consumer applications. But here's what I found - MOE-issued AI tutors don't teach up to major exams like the GCE A-levels, apps like Penseum are basic info synthesis tools that rely on your provision of notes, and even really great institution-ready startups like Pallo have a definitive response cap that protects their bottom line.

Response capping is the most obvious blocker of learning for myself - I'm not able to request deep dives when I want to. But the most fundamental issue is that every learning journey, content that is taught by an LLM wrapper, exists in a closed system.

What if I wanted to double check the sources of what chunk is being pulled from an application's proprietary RAG? What if I wanted my historical learning data, style, focus metrics to follow me to my notes, other study tools, etc?

Every current edtech incumbent is forced to protect such layers because opening up would mean killing their moat, and equivalently the favor from their VCs.

Build-in-public is the second principle. The strategic log is open. The architecture is documented. The roadmap, the unit economics, the pricing logic, the vocabulary discipline around what counts as proprietary IP and what doesn't - all of it is written down and available to anyone who wants to read it. The reason isn't transparency-as-virtue. It's that a tool built for serious learners earns trust by being inspectable, the same way the chunks in every answer are inspectable. Verdict is what it claims to be because you can check.

§ Follow the build

The build is public.

Discord is where the build conversation happens. Follow on socials for shipping updates, strategic shifts, and the occasional thread on why a decision went the way it did.

Join the Discord ↗Follow on X ↗