Skip to content


Working With The Eyes Of AI

AI, Computer Vision5 min read

Sayantani Basu

Meet Sayantani Basu

AI Engineer @ Wells Fargo

Bengaluru, India

Sayantani Basu is an AI Engineer with a specialization in Computer Vision. She worked on creating ML models to create Facial Keypoint Detection, Image Captioning, and Medical Imagery.

Sayantani has 5+ years of experience in Computer Vision and has been working in the IT industry for the last 11.8 years. She has an interest in autonomous systems.

What inspired you to pursue a career in Artificial Intelligence?

I have worked as a software developer for a major part of my career.

My manager in my last team where I was working as a software developer, was writing a book on AI. He would often carry around his manuscripts in his bag. And one day he asked me to have a look at one of his chapters.

When I started reading the content, I was so intrigued that I had to go back home and start finding out as much about AI and its capabilities. It did not take much longer to get hooked onto it.

AI is like magic, the possibilities of using AI to affect people’s lives are immense.

From rescue drones dropping food and medical supplies to creating applications that are able to diagnose diseases at par and often better than the most skilled medical practitioners, the things that we can do with AI are immense.

That is a great purpose for my life to work in the field of AI.

What are the function and implementation of Computer Vision in artificial intelligence?

Computer Vision is a field of AI that enables a machine to mimic the human Vision system. The human vision system is highly complex, and we still do not fully understand how concepts like perception work in human vision.

An example of perception is where we can recognize a person we might have seen a long time back and maybe at a time when they looked slightly different, but we are still able to recognize them.

We do not know what and how the brain stores the information in a person’s face that we are still able to recognize them even after certain changes in their facial features due to age or other factors.

But though the field of Computer Vision is still under a lot of research, we have made huge breakthroughs in this field in the last decade.

A very potent example of the implementation of Computer Vision would be in healthcare. It has been seen that in certain areas like in Detection of Skin Cancer a machine was able to achieve higher accuracy than a human medical practitioner.

What happens here is, skin cancer symptoms are very similar to acne-prone skin.

And it takes a trained eye and going through multiple skin samples to predict with accuracy if a certain sample of skin exhibits cancer or harmless acne.

When a computer vision application was trained on multiple samples of skin with and without cancer-like symptoms, it was able to predict with higher accuracy than a doctor, if a certain sample of skin is cancerous or shows some other signs of skin disease.

Why AI and deep learning are useful for object detection?

As an AI Analyst, my job role is to get the data which comes on chatbots and then train the chatbots again so that it can understand more things.

It is just human in the loop of AI thing. I have to train the Algorithm again by the real-time data on which our machine was unable to reply earlier. 

What kind of analysis do AI analysts specifically work?

Object detection involves detecting specific features in a particular type of object.

For example, if we are trying to detect if Horses are present in a particular image or not, we have to teach our system to learn the specific features of a Horse.

In a Deep Learning network, when we feed the network with multiple images with and without horses for the purpose of training the network, the network is able to extract features on its own that it identifies with the presence of a Horse.

We do not have to explicitly mention the features that identify a Horse. The deep learning network is able to do this entirely on its own.

So, after training the network when we give it an image of a Zebra, which is shape-wise very similar to a Horse, the network is successfully able to predict that the image does NOT contain a Horse.

This is how powerful a deep learning network can be.

Do you have any Ai project that you are working on and want to share your experience with us?

I am currently in a very initial stage of building from scratch a Drone that would have the ability to navigate through closed spaces on its own and will be able to detect objects and also recognize faces.

My intention is also to give it the power to understand and interpret language. In other words, it would have capabilities of a ChatBot but also some emotional intelligence.

Some open-source frameworks are available for the implementation of this project, but the interesting and challenging part for me would be to integrate all this in an autonomous drone and explore the possibilities of such a product.

Mention any industry where Ai could have a big impact? and Why?

I believe the field of autonomous vehicles is going through a great phase and things will only get better with more advancements in AI.

We already have the Self Driving Cars which was a huge disruption. I think in the field of Flying Vehicles it will have a huge impact, which is going to change the way we live.

As in the case of Drones, which could have multiple applications from retail to security to hospitality.

Flying Cars, which is also my personal dream when successfully established will not only transform the way we commute but also the way we live.

Where do you see yourself in 5 years in the field of AI?

In the next five years, I hope to have a Company of my own which will have a Research Lab that can come up with products in AI that would aid in better healthcare facilities.

Also, I want to build a Flying Pod which would be Solar Powered and will cause a disruption in our commute system.

What role models have been your greatest inspiration to get into AI?

One of my greatest role models has been Andrew Ng. When he famously says “AI is the new Electricity” that kind of encompasses it all and brings a burning desire to be able to implement applications in AI in all fields possible.

One of my other recent role models has been Parul Goel, who has recently become a Kaggle Notebook Grandmaster.

Being a new mother, I have taken a lot of inspiration from her, from her posts where she mentions how she transitioned into the field of Data Science during her maternity leave and how scary the job descriptions have been for a new entrant to this field.

Every person’s journey in this field is a new struggle which can be an inspiration for others around!

For more about Sayantani Basu projects go to her YouTube Channel and Medium

A Book Sayantani Basu recently reviewed: The Computer Vision Workshop