Last week, together with our colleague from UPSI, Dr Raziff, I had the opportunity to pitch AINNA at MTDC. For me, the most interesting part wasn't the pitch itself, but the discussion around a bigger question – how do we actually move AI from research and experimentation into something that can be used in real operations?
We shared how AINNA is building NeuralOps and AI agents based on real operational needs. Not every task requires a large model. Some tasks can be handled by rules, parsers, smaller models, or even fully detached systems. The goal is straightforward: make the system more reliable, cut unnecessary token usage, reduce hallucination, lower compute costs, and make AI practical enough to run inside real businesses.
This is also where collaboration becomes crucial. Researchers bring knowledge, validation, and new ideas. Industry brings real problems, real workflows, and real operational challenges. MTDC can play an important role in helping technology move beyond development into commercialisation and wider industry adoption.
For me, Malaysia already has strong talent, researchers, and technology builders. The challenge is how to connect all of these capabilities and turn them into solutions that can operate reliably, scale effectively, and create real commercial value.
It's still early and there's plenty for us to prove. But discussions like this with MTDC, together with our fellow Dr Raziff from UPSI, help us see the road ahead more clearly – build locally, validate with real use cases, commercialise properly, and eventually scale beyond Malaysia.
#AINNA #MTDC #UPSI #NeuralOps #ArtificialIntelligence #AICommercialisation #AgenticAI #DeepTech #MalaysiaAI
Ruang pembaca
Apa pendapat anda?
Komen baharu dihantar untuk semakan terlebih dahulu. Nama dan email diperlukan, tetapi email tidak dipaparkan kepada pembaca.
Tulisan yang bagus. Dr raziff sahaja dah berbaloi.
commercialise properly, and eventually scale is the part I would forward to my boss.
Sent this to two people already. cut unnecessary token usage, reduce is why.
Bookmarked, mostly for parsers, smaller models.
I read this twice. Last week, together with our is the part that stuck.
The figures on lower compute costs, and make make more sense than most posts. Need to read this part again.
First piece that handles scale effectively, and create real honestly.
Setuju soal dr raziff, tapi eksekusinya tidak mudah.
Clearer than the vendor decks I get about researchers, and technology builders.