Photo by Anne Nygård on Unsplash

One of the biggest challenges in data science is to be able to quickly deploy a model in production. And, the story doesn’t end here. Actually, the dev-deploy interlock has been long desired to be very simple, flexible and always plausible as we need to update and scale our models frequently and the changes are needed to be available in production immediately and seamlessly.

The motivation for penning down this article was my struggle with making models deployable while I was researching on on-wire ML algorithms and they needed be deployed for consumption in a networking compute environment during 2017–2018.

Photo by Road Trip with Raj on Unsplash

When I was a kid, it was like I wake up everyday with new superpowers. It seemed so magical..someday, my hands would have grown stronger, other day, I had started running. One day, I got biting power through a new pair of teeth. Each day I was only gaining powers. I must have been so curious and excited and every few days, I must have been doing so many new things, having so many first time experiences. Years later, I was a fully grown and supposedly a powerful human being. But, I could no more think how many novel and…

Image courtsey:

About twenty years ago, there was a convenience store named Ram Lakshaman in Dandra, India. A ten year old girl would often go to that store with her mother and pass a grocery list. The store owner, a middle aged man known as Kush uncle in the locality, would ask a teenage helper to find and bring those items on a front wooden table. Kush uncle would then weigh the finalized items and put them in a bag brought by the girl and her mother. It was rare that Kush uncle would have all items in his store. And he…

A Live Visualization of the current global state of Coronavirus Infections

Image source: Flickr

Whether you are an enterprise or a small business striving to make an impact on the market, you need to understand the sales, market trend, your revenue drivers and much more. Plotting charts on an excel spreadsheet is old school. A well-organized dashboard is what today’s businesses need to own. The only downside is that all tools available in the market to achieve this quickly has a cost. And, of course, many of us want a free way to achieve this at least initially during the exploratory data analysis phase. I am addressing the same need here and providing a…

By Unknown — Tokyo National Museum, Emuseum, Public Domain,

In Tibetan Buddhism Hungry Ghosts (Tib. ཡི་དྭགས་, Wyl. yi dwags, Sanskrit: preta) have their own realm depicted on the Bhavacakra and are represented as teardrop or paisley-shaped with bloated stomachs and extremely thin necks to pass food such that attempting to eat is also incredibly painful. So, these ghosts have bellies as vast as mountain valleys and mouths like the hole of a needle. Even if they find food or drink, they cannot consume it. Thus they suffer from hunger and thirst. …

Textual analytics on conversational transcripts using Keras-Tensorflow-LSTM, Watson NLU and Watson Personality Insights.

Photo by Jonas Jacobsson on Unsplash

I recently talked about deriving attributes from sitcom transcripts and determining the possibilities of learning the essence of making a popular television show at Data Science Salon, Los Angeles. This article would present the investigation which went in and what were the results I got post training on neural networks as well as leveraging pre-trained Watson NLU and Personality Insights models. The most challenging part of the journey was to prepare data by randomly searching for public web transcripts of these shows on the internet.

Data Science Salon, LA 2019 talk on ‘Can Data Science help us find what makes…

With Working Example(s)

Semantic segmentation and alpha blending to change the background of an image. Please seek permission before reusing this image by emailing to

Image courtesy:

In this world of ever increasing data at a hyper pace, we use all kinds of complex ensemble and deep learning algorithms to achieve the highest possible accuracy. It’s sometimes magical how these models predict, classify, recognize, and track unknown data. Accomplishing this magic and more has been and would be the goal of intensive research and development in the data science community.

But, around all this great work, a question arises:
Can we always trust this prediction, classification, recognition, and tracking?

A variety of reasons, like lack of data, imbalanced datasets, biased datasets etc. can impact the decision rendered…

Image courtesy —

If you are planning on opening a physical store for selling consumer goods, you probably need to give it another thought. With the digitization and online shopping available at our fingertips, the option of going to the store looks way more time-consuming and inconvenient. You might ask, if this is so, how is Costco flourishing with a billion dollar profit? I have an answer to this — look at Costco’s business model. It does not try to confuse you with an array of options in different categories. Rather, it does the work of qualitative filtering and you just need to…

I love startups and how there is no bound to what one could do to make this world a better place. And, since I work in Data and AI domain, I was more interested in exploring the Data and AI startups on AngelList. Gathering data from AngelList was not trivial as it would not let you export more than a 100 records at a go. So, I applied multiple filters, like Type=”Startup”, Market=”Big Data Analytics”, “Data Science”, “Big Data”, “Machine Learning, Predictive Modeling”, “Computer Vision”, “Natural language Processing”, “Neural Networks”, “Deep Learning”, “Reinforcement Learning” and then excluded the duplicates later…

Shilpi Bhattacharyya

Writing is therapeutic. Data is the lake where I fish.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store