FildraAI is built in Taiwan, where our team knows AI, and validated on real farms and in real markets in Zambia. We put the two together so a farm can keep its money straight.
What we're fixing
Money comes in and goes out, and nobody writes it down. The farm owner is often far away. By harvest, no one can say where the money went. That is what FildraAI fixes.
Our mission
To turn what a farm does into money it can see, and to make what it learns one cycle pay off the next. Keeping the record is not the point; the decision it leads to is.
The Problem
A bag of fertiliser costs real Kwacha. A worker is owed real Kwacha. If no one writes it down, the money slips away and nobody knows where.
The owner often lives far off and pays for the farm from there. The farmer is in the field with a phone and a patchy connection. Between them, the truth gets lost.
FildraAI fixes this. You or your team speaks what happened. Field State does the books, and you can share one clear weekly picture, with proof, with whoever funds the farm.
A chat AI can give a smart-sounding answer and still know nothing about your farm. It does not know what you spent, who you paid, or who still owes you. So it cannot tell you if the farm is making money.
A Story Across Two Places
The reason for FildraAI comes from Zambia. The skill to build it came from Taiwan. We put the two together so a farm can keep its money straight on a normal phone.
Zambia · Southern Province
Our founder, Chiuzu Chilumbu, grew up in Zambia. Here a maize farm feeds a family and pays the bills. He saw the same trouble again and again. Money came in and went out, and nobody wrote it down. By harvest, no one could say if the farm had made money or lost it.
Taiwan · Agricultural Region
Chiuzu studied at Tunghai University and worked on the campus farm. Then he did a master's in AI at National Tsing Hua University. That is where he learned to build the tools behind Field State. It is also where one question stuck with him: how do you build AI a farmer can actually trust, not a clever box that hides how it works?
What the technology does
Everything we build answers one question: does this help the farmer who is farming blind? Here is what the technology does.
Our Team
FildraAI is a small team: research, design, and people who know the markets in Africa and Asia. This is not one person's show. It takes all of us to build something farms can really use.
Founder & Research Lead
Leads the AI work and keeps Field State focused on one job: show the farm's money clearly, with proof.
Education
M.S. Computer Science & AI National Tsing Hua University, Taiwan
Experience
ML/AI Research & Development Engineer Author, DenseSwin (IEEE 2024) · Field validation across Zambia & Taiwan
Graphics & Design
Shapes the visual language, product presentation, and communication assets that make the platform easier to understand.
Education
M.A. Visual Communication Design National Taiwan University of Education
Experience
3+ years in brand & product design
Asia Market Analyst
Tracks market needs, partner opportunities, and regional signals across Asia to keep product strategy grounded.
Education
M.S. International Business and Political Science National Chengchi University
Experience
5+ Asia-Pacific market research & partnerships Firm grasp of regional dynamics
Africa Market Analyst
Supports African market research, farmer context, and validation priorities for practical deployment.
Education
B.Sc. business and Finance University of Lusaka
Experience
Smallholder programs & field research, Zambia Extensive experience in agricultural development
How we work
Fildra meets farmers as a practical tool: speak what happened, see your money, know what changed. The research that makes this possible sits behind the product, and earns its place only when it makes that tool better. A farmer should never need to understand our AI to understand why Fildra is useful.
Audio-first farm records, money you can see, and questions answered from your own history. This is what we build, what we sell, and what we judge ourselves on.
Computer vision, audio and language models, and agricultural knowledge work. It broadens the platform once real usage and revenue justify it, not before.
Farmers recording week after week, accurate money from a voice note, a remote owner paying for visibility. Research earns its priority only when it improves the product, supports a real partnership, or builds a lasting advantage.
Field Validation Work
"Utilizing a DenseSwin Transformer Model for the Classification of Maize Plant Pathology in Early and Late Growth Stages: A Case Study of Its Utilization Among Zambian Farmers"
IEEE PublicationAn AI model that looks at a photo of a maize plant and spots signs of disease, early on and later on, so a farmer can act before it spreads.
The research behind it
Chiuzu's master's research asked a simple thing: can a phone photo tell a Zambian farmer their maize is sick?
He built a model called DenseSwin to read those photos. The aim was to catch disease early, while a farmer can still save the crop.
In the study, the model got it right 97 times out of 100. That showed AI can really help on a farm, if it is built with care and tested for real.
But the research taught us a bigger lesson, and it changed what we build.
What Made the Difference
Money in. Money out. Who still owes you. Those are facts you can write down and add up. A farm runs on those facts, every single week.
So we built the books on simple math you can check. Field State does not guess your profit. It adds what came in, takes off what went out, and shows the working so you can follow it.
And the farmer speaks in their own language, any of our supported languages. They say what happened. Field State turns it into the books.
Your money is your money. You should not have to trust a machine blindly. If Field State cannot show how it got a number, you cannot check it. So Field State always shows the working, and keeps your farm's record so last week is still there next week.
What the Books Track
What We Built
Field State is the farm app, and we are building it now. You or your team speaks what happened. Field State does the books, keeps the farm logs apart from the money, and tells you what needs you, from your own record. You can share one clear weekly picture, with proof, with whoever funds the farm.
The farmer says what happened in their own language. Field State turns it into the books. No forms to fill in.
Money in, money out, who still owes you. Simple math you can check, with the working shown.
Everything is recorded once, kept straight, and never rewritten. What came in, what went out, what is left, with proof on every line.
Designed for real field conditions, even when the connection is patchy. Speak now, and it syncs when the connection returns.
Speak in any of our supported languages: Nyanja, Bemba, Shona, Swahili and more. Field State keeps up.
Field State reads your own logs and shows you what is slipping: a debt unpaid, a cost with nothing to show for it. From your record, not a stranger's guess.
We are working with our first 10 farms right now, owners and farmers, getting the books right week by week. Recording and the weekly report are free, and payments aren't live yet. We are not here to be one more AI app. We are here to show your farm its money, with proof.