Anti-Scam Bangladesh

01 Spotting the Window

My third project is an anti-scam and cyber-awareness platform in Bangladesh, built with my closest friend — a Bangladeshi — and aimed at government and institutions (B2G / B2B). The surface reason is simple: he's local, he knows the ground, and we're a local team. But what actually made me commit is that this market is structurally mismatched.

Almost all cybersecurity education today is built for people with an IT background — CTFs, red-vs-blue exercises, enterprise security training. We're doing the opposite: building for ordinary people with no IT background. Bangladesh is developing fast and smartphones are already widespread, yet the matching anti-scam education is essentially nonexistent, and ordinary people are harassed by fraud constantly. It's a badly overlooked market that is genuinely being hurt.

Why Bangladesh, of all places? The full argument is its own piece — I wrote it up as YVI, my decision framework — but the short version is arbitrage: at the same density of ability, I'm roughly one standard deviation out in the US or Singapore, two in China, three in Malaysia — and in Bangladesh, four. The same ability is worth wildly different amounts in different markets, and Bangladesh pairs my widest usable gap with a national cycle still on its way up.

02 The Bet, and What I Traded For It

Traditional anti-fraud work runs on classes, training sessions, and flyers. The process is dull, and ordinary people won't engage with it. It's fundamentally against human nature: it asks you to be serious, to concentrate, to actively study. We go the other way — working with human nature. People already want to be entertained, to scroll their phones, to play a couple of rounds, so we hide the anti-fraud message inside what they're already doing and let them build immunity while they play.

In product terms, that becomes three lines: a scenario-simulation game, a set of low-friction mini-games, and AI short dramas. We're not after polish; we're after reach. The hard part of anti-fraud has never been that the content isn't good enough — it's that ordinary people simply won't spend time on it. Every notch of friction loses another slice of people. Only by pushing the barrier close to zero, in formats locals are already playing, can we pull in even the people who would never attend an anti-fraud class. Herd immunity comes from breadth, not depth.

03 Building It

The three lines split into two layers: an outer layer that reaches broadly and pulls people in, and a core layer that does the real, in-depth training.

The outer layer is the mini-games and the AI short dramas — two parallel acquisition lines, both distributed on mass platforms like Facebook, TikTok, YouTube, and Instagram, competing for attention with low-friction, entertaining content. The mini-games exist to guarantee fun and spread, drawing people in while injecting basic cyber-safety knowledge and anti-fraud awareness as they play. The AI short dramas carry the high-frequency reach — essentially a "scam vaccine" for the public: the same scam setups appear over and over until "I've seen this trick" becomes a reflex, building herd immunity at a national scale.

The core layer is the scenario-simulation game, and this is the real heart of the platform. We turn anti-fraud and cyber-safety training courses into playable, engaging simulations where the user walks through a scam end to end — making choices, falling into traps, and seeing the consequences firsthand. It also records how a user changes before and after training and produces concrete numbers, so the effect can be measured. This is the core product we actually deliver to governments and institutions; the outer-layer mini-games and short dramas are, in essence, funnels that feed it.

I personally handle the AI short dramas in the outer layer, and I've turned them into a pipeline I can replicate quickly: write the script, generate a storyboard with GPT image-2, fine-tune each frame until it's right and compress it, then hand it to Gemini Omni to generate the video. The path from script to finished clip is highly automated, so when a new scam appears, we can translate it into a finished piece in very little time.

Connect that pipeline to distribution and you get a complete loop: the government and police continuously supply first-hand scam intelligence, the games and dramas translate it into mass content, and short video pushes it out. In essence: scam intelligence → mass content → herd immunity. The longer this loop runs, the deeper our understanding of the local fraud ecosystem, the larger our content library, and the wider our distribution network become — hard for a latecomer to catch up with in a short time.

04 Seeds for the Future

I'm walking this in three stages.

• Short term — make the anti-scam outreach itself solid and get that loop running stably.
• Medium term — partner with the Bangladeshi government on broader cybersecurity training, extending from public-facing anti-fraud awareness into security capacity-building at the government and institutional level.
• Long term — bring in the Chinese resources behind me: some of the world's best engineers and the strongest manufacturing and supply chains. Using anti-fraud and digital-safety education as the entry wedge, I'd move step by step into more cybersecurity verticals — from consumer-level awareness education toward enterprise-level and infrastructure-level security services.

For me, this is more than a startup project. It's the first real-world deployment of YVI, my decision framework: (personal-capability arbitrage × country-cycle arbitrage × industry-cycle arbitrage) ÷ risk. And the conclusion is clear — picking the right market matters far more than simply grinding harder on ability.