Welcome to Deepak’s home on the Web.
I specialize in machine learning and software engineering. Jump ahead to my bio.
Recent Work
What I have been up to lately:
- (June 29 2021) Analyzed hailstorm data to find safe places for solar panel installation in the US
- (June 17 2021) Reduced program runtime by 10x by replacing blocking I/O with a thread-pool
- (June 4 2021) Did a book review of Futureproof, capturing what I liked about it
- (May 7 2021) Found that radar data and deep learning can detect California irrigation with 95% accuracy
- (May 6 2021) Did a review of some books in the realm of psychology and psychotherapy
Most Popular
Here is the most popular content.
It always amazes me how these remain popular, because they are so obscure, and in my mind, so ancient.
- (Oct 21 2016) A technical post about closing a network connection in one shot from your application
- (Feb 1 2017) An obscure post about what to do if disk is full on a XenServer hosting virtual machines
- (Feb 2 2017) A race condition can occur in Python if you don’t create a directory correctly.
- (Aug 5 2020) A post on how we built a global ML model with Earth Engine
- (May 21 2016) A tale of debugging a bad error message. This post is probably popular more because of the error message than what I wrote there.
My Favorites
Here are some of my personal favorites:
- (2020-2021) Our journal article on global irrigation extent and lightning-talk at Google Geo-for-Good summit
- (June 28 2020) We can take a concept from functional programming to apply in real-life: lazy-eval
- (June 10 2016) When I found a bug in the FreeBSD operating system
- (April 10 2021) I very much liked the novel by Julian Barnes, Sense of an Ending
- (April 30 2021) I had fun building a Lego grand piano
- (2007-2008) I selected and translated some verses from a philosophical work in my native tongue, Kannada, into English
Bio
At work, I develop end-to-end machine learning solutions. I have experience with data pipelines, distributed parallel programming systems, cloud platforms, and model operations. I can work with high-level programming languages like Python as well as lower-level languages like C and (x86) assembly. Recently (2020) a graduate of Data Science at UC Berkeley, I also hold a Master’s degree in Computer Science from Ohio State (2008) and a Bachelor’s degree in Computer Science from PESIT, Bangalore, India (2003). I have worked in 3 startups and 2 large companies, in Bangalore and SF Bay Area. I currently live in Mountain View, California.
During my free time, I do pro-bono machine-learning research for sustainability with faculty at UC Berkeley. I have experience with large-scale machine learning on geospatial data including optical, multispectral and SAR remote-sensing data, regression and classification models, time-series analysis, and deep learning. I recently published a research article on global irrigation extent, and gave a talk at Google Geo for Good summit. My data-science portfolio.
Full profile, including recommendations, available on LinkedIn.
You can follow me on Twitter.
Recruiters, please note: I respectfully refuse to do “Leetcode” interviews. I am interested, however, in a real-world problem in your company and seeing whether I can help you with it.
Meta
This website has been around since 2000, although it has evolved almost continuously with technology… and me.