Interview with Kritika Prakash
Where are you based?
"I’m currently based in Bangalore, India."
What do you do?
"I lead the Differential Privacy Research Team at OpenMined. I am an MS by Research student in Computer Science at IIIT Hyderabad (India) where I work on building algorithms for power trading in smart electricity grids. I am also an artist."
What are your specialties?
"I come from a research background and work on interesting problems in Deep Learning, Differential Privacy, Reinforcement Learning and Game Theory. A lot of my work involves Python and Java programming, problem formulation, technical reading and writing. Currently, I am working on a research project at OpenMined to automate Differential Privacy analysis for Deep Learning."
How and when did you originally come across OpenMined?
"I first came across OpenMined 1.5 years ago. In 2019, I started working on a side project on Differential Privacy. I faced some technical roadblocks and started looking online for implementations of Differentially Private Machine Learning Algorithms. That’s when I came across OpenMined and explored PySyft as well as the blogs.
Soon, I filled out the survey for the mentorship program to get mentored and more involved with the community."
What was the first thing you started working on within OpenMined?
"I became a mentor as a part of the Mentorship Team in November 2019 and had a lot of fun helping other beginners figure out how to contribute to OpenMined and achieve their own goals."
And what are you working on now?
"I am working as a research scientist on the project “Automatic Sensitivity Analysis for Differentially Private Deep Learning”. As part of the Differential Privacy Research Team, I collaborate on fun research-based projects on different applications of Differential Privacy to the real world, and how we (OpenMined) can build tools to serve this purpose. I also give talks and write blog-posts on Differential Privacy."
What would you say to someone who wants to start contributing?
"Start small and ask for lots of help. It’s always confusing and overwhelming at the beginning. The OpenMined slack is your best friend here. This community is incredibly helpful, friendly and resourceful. Take initiative. Look for things to do/fix which align with your interests, and then, just give it your all. Once you’re familiar with the basics, the best way to contribute is to start working on a project, and learn all the skills in the process. Contributing to open source is a great way to work on your self-growth and build skills in various areas such as coding, teamwork, research, communication and more."
Please recommend one interesting book, podcast or resource to the OpenMined community.
"Lex Fridman’s series of podcasts on “Artificial Intelligence” have been super helpful in making me feel connected to the most thriving parts of the computer science global community. They’re easily accessible on YouTube. While textbooks help you get strong technically, such interviews give you a perspective on how different fields connect with sub-fields of computer science, and what role we play in shaping tomorrow’s world as engineers and researchers.
Another resource I’ll strongly recommend to beginners in Machine Learning is fast.ai’s courses on practical Machine Learning. The content is taught in a top-down fashion with a lot of jupyter-notebook style of experimental coding along the way. It really helps you get a practical understanding of what’s happening and solidifies your learning in the process."