Photo by Isiah Gibson on Unsplash

Arjun tapped away at his keyboard, changing a few model parameters, and started the simulation. He was testing out a new powertrain configuration, the result of months of research. The electric vehicle startup he had joined a year ago had ambitious goals, but he liked his work and the general company culture. He was surrounded by people smarter than him, and learnt a lot every day.

He waited for a minute and glanced at his watch — 5.56 pm. He had dinner plans soon, and it would take a while to get there in Bangalore traffic so he’d need to…


Statistics is used for different purposes in different contexts. For example, descriptive statistics is the branch that deals with summarizing an existing sample of data. Descriptive statistics deals with the sample itself — there’s no need for any extrapolation. However, a more challenging problem that we have to address regularly in the real world is using the sample to infer something about the population from which the sample was drawn. For example, let’s say you rolled a die 10 times, and saw these results:

4 5 2 6 1 5 6 5 3 1

Descriptive statistics can tell you that…


If you look at published research papers in fields involving data analysis (which nowadays is basically every field), you’ll often see a little asterisk* next to the numbers with a footnote saying “statistically significant”. Statistical significance is essentially a check that we can do to verify that our results aren’t due to chance, and they involve computing something called a p-value (probability value). And while the statistical significance asterisk is often interpreted as a stamp of approval for the validity of the research, there are some issues we can run into using p-values blindly. …


An ubiquitous statistical phenomenon that affects our everyday lives, but which most of us are unaware of.

The LA Dodgers, Los Angeles’ resident baseball team, appeared on the cover of Sports Illustrated’s August 28, 2017 issue following a run of tremendous form. They had won 71% of their games and were on pace to tie the record for the most wins in a season. The cover came with the caption “Best. Team. Ever?”

The team then went on to lose 17 of their next 22 games, and would eventually lose in the World Series to the Houston Astros. …


70% — That’s the percentage of people who have felt an episode of impostor syndrome at some point in their careers. Dr. Pauline Rose Clance was responsible for the study that published this number, as well as a lot of the initial research into impostor syndrome. She defines it as “the psychological experience of believing that one’s accomplishments came about not through genuine ability, but as a result of having been lucky, having worked harder than others, or having manipulated other people’s impressions”.

We’ve all been there at some point. The feeling that you’re underqualified or don’t deserve to be…


What exactly is reinforcement learning, and how does an RL algorithm work in practice?

In 2014, Google acquired a British startup named DeepMind for half a billion dollars. A steep price, but the investment seems to have paid off many times over just from the publicity that DeepMind generates. ML researchers know DeepMind for its frequent breakthroughs in the field of deep reinforcement learning. But the company has also captured the attention of the general public, particularly due to its successes in building an algorithm to play the game of Go. Considering the frequency with which DeepMind makes progress in the field — AlphaGo, AlphaGo Zero and AlphaZero in the last couple of years…


After spending 1.5 years in an MS programme at CMU, I’ve decided to move into an applied ML role in industry instead of continuing on to a PhD. I’ve had a lot of discussions with people about the case for and against a PhD, and I thought it’d be useful to explain my thought process in deciding definitively against a PhD. …


Last week, from June 24th — 26th, Google held its Google Cloud Next ’18 conference at the Moscone Center in San Francisco. It was Google’s biggest event ever, with 25000 people attending.

These kinds of specific product-related conferences often end up becoming a sales pitch for the product (justifiable considering the kind of money Google likely put in to host it), so I won’t talk too much about the keynote or the individual talks. There was however an interesting theme across the conference of even greater investment into AI for Google Cloud. This is likely the one major distinguishing factor…


In the previous post, I gave an intuitive view of the workings of reinforcement learning, and worked through how the learning process would play out for tic tac toe and chess. Now, let’s get to the really interesting stuff — Go!

Go is a Chinese board game that’s been around for millennia. The rules themselves are fairly straightforward, much simpler than chess. Players take turns putting black and white pieces on a 19x19 grid, and the one who manages to surround the most territory wins. …

Deepak Dilipkumar

Machine learning engineer

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