As the Pembangun Sistem AI at AINNA, I am currently helping prepare two pitch sessions at MOSTI Cyberjaya. Both builds started with the same engineering question we keep asking in the lab: how can technology remove friction instead of adding more?
1️⃣ Penyata Kewangan for Street PKS
Most street-level PKS do not skip financial statements because they do not care. The existing process is usually too complex, too costly, and too slow, so entrepreneurs keep operating without structured records.
We approached it from a systems angle. Rather than forcing users to learn an accounting system, we designed a pipeline that learns how entrepreneurs already work.
✅ Phase 1 is shipped and running.
https://lnkd.in/gfNKAXHt
The pipeline converts bank statements into Excel automatically and is free to use. One architectural decision matters more than the feature itself: AINNA’s servers never store uploaded bank statements. Fail are parsed in memory for format conversion only, then discarded. This is not just a privacy policy; it is a system design choice. Financial data belongs to the user, not the platform.
Phase 2 is now in development. We are wiring up automated classification and statement generation so the accountant-level output requires minimal manual input.
The real goal is not just speed. It is making compliant financial record-keeping accessible to street PKS that would otherwise be left out.
2️⃣ Sistem AINNA NeuralOps
While the market chases bigger LLMs, we are optimizing the infrastructure that runs them.
The Sistem AINNA NeuralOps is an end-to-end AI infrastructure platform made up of:
• Pelayan LLMs
• AI Orchestration
• Ejen AI
• Seni Bina Terpisah
Every component is demand-activated. Nothing sits idle burning GPU cycles, electricity, and cooling budget.
That design gives us:
• Higher effective GPU utilization
• Lower operating cost per inference
• Reduced electricity and cooling load
• A modular, scalable, and more sustainable AI backbone
The next phase of AI will not be won by model size alone. Smarter architecture—routing, orchestration, detached compute, and agen lifecycle management—will decide what is actually deployable at scale.
From helping street PKS close their books...
...to running enterprise-grade AI infrastructure...
AINNA’s mission has not changed. We build systems that lower digital barriers, simplify compliance, and put advanced technology within reach of more people.
Not everyone will buy into the approach. But if engineering choices can widen participation and create real opportunities for entrepreneurs, they are worth a serious look.
We hope these two systems can contribute something useful to Malaysia’s innovation ecosystem through the MOSTI platform.
🇲🇾🤲
#MOSTI #Cyberjaya #Ainna #NeuralOps #ArtificialIntelligence #StreetSMEs #FinancialStatements #EnterpriseAI #LLM #AIAgents #DigitalTransformation #MalaysiaTech #Inovasi
