As a child growing up in the bustling streets of Satkhira, Bangladesh, I often witnessed the stark realities of healthcare access. Long queues at rural clinics, the seasonal dread of dengue outbreaks, and families struggling to get timely medical advice were not just news headlines they were lived experiences.
These observations, etched in my memory, ultimately fueled my passion for artificial intelligence (AI) and its potential to transform healthcare in resource constrained settings like our own.
Today, as an independent researcher specialising in trustworthy AI for healthcare, my work is driven by a singular question: How can we make AI a reliable, fair, and accessible partner in medicine, especially for communities that need it most?
From Jessore to global labs: A journey rooted in local need
My undergraduate academic journey began at Jashore University of Science and Technology, surrounded by the vibrant energy of Bangladesh’s emerging tech scene. This foundation propelled me to advanced research in the United States, where I delved into multimodal deep learning, exploring how AI can interpret complex data from biosignals, images, and texts to predict diseases.
To the next generation of Bangladeshi engineers and scientists: the world of AI needs your perspective. We need to build more than just accurate algorithms; we need an ethos that prioritizes fairness, privacy, and interpretability
Yet, my focus has always circled back to Bangladesh its health challenges, resilient people, and untapped potential for AI-driven change. My professional identity as an “enthusiastic data scientist and machine learning specialist… passionate about developing trustworthy, interpretable, and robust deep learning methodologies” is not just a title; it is a mission statement born from local context.
Building tools for Bangladesh: Dengue, data, and decision support
One of my key projects directly addresses a recurring national crisis: “AI Chatbots for Dengue Symptom Triage in Bangladesh.” Drawing from my experiences volunteering with organisations like the Red Crescent Society in Jashore, I developed a chatbot system that uses natural language processing to provide instant symptom assessment in Bengali or English. In rural areas where doctors are scarce, such an AI assistant could guide people to seek help early, potentially saving lives during an outbreak.
Beyond acute crises, understanding long term trends is vital. In another project, my collaborator and I used data science tools to perform a “Comparative Analysis of 2019 and 2023 Health Bulletins” of Bangladesh. By applying machine learning to national data on maternal health, vaccination, and non-communicable diseases, we can uncover critical trends to inform smarter public health policy. This work exemplifies how AI can turn our own national data into actionable intelligence for a healthier future.
The core of trust: Fairness, privacy and explainability
For AI to be truly adopted, it must first be trusted. This principle of “Trustworthy AI” is the cornerstone of my research and is deeply relevant for Bangladesh.
• Fairness Aware AI: My work on fairness in ECG based disease prediction ensures AI models do not perform poorly for any demographic group. In a diverse country, building equity into medical AI from the start is a moral imperative.
• Privacy Preserving Collaboration: My “MedHE” framework allows hospitals to collaboratively train powerful diagnostic AI models without ever sharing sensitive patient data, using a technique called federated learning. Imagine a secure network where hospitals in Dhaka, Chattogram, and Rajshahi could improve a shared AI for heart disease while fiercely protecting every citizen’s privacy this is the future we are building.
• Beyond the Black Box: I research “explainable AI” so that when a model makes a prediction, it can also show why. For a doctor in a busy upazila health complex, this transparency is the difference between blind faith and informed decision making.
The independent path: Carrying Bangladesh onto the global stage
Choosing the path of an independent researcher has allowed me the freedom to focus on problems that matter most and to bridge my Bangladeshi perspective with global scientific discourse. This journey, from organising events with the Bangladeshi Student Association in the US to now publishing at international conferences, has taught me that impactful science requires both technical excellence and a deep connection to human need.
The upcoming feature on my work in a national media outlet is a humbling recognition that research grounded in local context can have global resonance.
A call for a Bangladeshi AI ethos
To the next generation of Bangladeshi engineers and scientists: the world of AI needs your perspective. We need to build more than just accurate algorithms; we need an ethos that prioritises fairness, privacy, and interpretability. Our country’s complex challenges are the perfect training ground for developing robust and ethical AI.
As I continue my work, I carry the spirit of Bangladesh its communal strength, urgent challenges, and immense potential into every line of inquiry. The goal is clear: to contribute to a future where AI is not a distant, opaque power, but a trustworthy tool for health and empowerment, built for the world by a community that proudly includes Bangladesh’s unique voice and vision.
* Farjana Yesmin is an independent researcher & machine learning scientist specialising in Trustworthy AI, Federated Learning, and Healthcare Applications. She is a graduate of Jashore University of Science & Technology and University of Tulsa, USA, focusing on making AI fair, private, and explainable. Explore her research at https://farjana-yesmin.github.io/