Intern Diary : Vijay Veerabadran

This blog post summarizes my experience as a research intern at Artifacia Inc. during the winter of 2016. As a Research Intern, I was asked to work on research problems relevant to the company at the intersection of Deep Learning and Computer Vision. My relationship with Artifacia began in August 2016 when I applied there for a winter internship.

Getting selected

I had been interested in deep learning since long, I had constantly spent a good amount of time and effort to learn the same. Once I was confident about my theoretical knowledge of the same, I started looking for places where I could apply it practically. This was when I got to know about Artifacia and how well it was performing in the AI/ML tech space. The selection process consisted of 3 rounds of intense technical interview. For the first round, I had to solve a machine learning challenge on Kaggle. Following which there was a telephonic interview which was completely about Deep Learning and its application to Computer Vision. This was followed up by one final round of interview, which was an algorithmic programming round. Upon performing well in these interviews, I was offered the role of a Research Intern and I joined them during December 2016.

Details about my research work

Considering my interest in generative image modeling, all my projects at Artifacia were aligned to this sub-domain. Some noteworthy projects I worked on include Image Super-Resolution, Image Inpainting, and Generative Adversarial Text-to-Image Synthesis. The stunning efficiency with which we were able to work on problems has to be owed to the systematic approach adopted by the team at Artifacia to solve problems, which I also started using. The approach can be summed up in the following three activities, performing a literature survey, identifying the top-performing solutions already out there, reimplementing some of them from scratch and performing a few architectural tweaks and changes to data handling based on experience and experimental results. Owing to this well-planned strategy, we got to work on multiple interesting problems in the short span of 3 months of my internship. I was given the freedom to work with tools that I was comfortable with, I worked on all my problems using TensorFlow.

The founders of Artifacia

My relationship with the founders at Artifacia has been and is an enlightening one. Having expected a bunch of managerial gurus as the founders, I was pleasantly surprised by the techno-managerial skills that they possessed. Right from having an in-depth interview about ConvNets to discussing work and ideas frequently now, I have been impressed by the founders’(Navneet’s and Vivek’s) comprehensive understanding of machine learning technology and how it can help with product innovation to solve real world problems.

Navneet who primarily looks after product and business at Artifacia is very skilled at arriving at solutions to real-world problems using Machine Learning/Deep Learning algorithms, which led to most of the problems I worked on being aimed at solving such problems. This quality of identifying problems that be addressed with is a trait I admire a lot in Navneet and he has been helpful in nurturing the same skills in me. I'm sure that it will be put to great use by me in the future. Navneet also loves to follow state-of-the-art research in AI and often shares what he likes with us. We still have regular conversations about my current projects and about possible collaborations. It has always been a pleasure working with Navneet.

I have been repeatedly stunned by the out-of-the-box solutions Vivek, who leads technology at Artifacia, gives to very tricky problems in the Computer Vision and Deep Learning space. Vivek has been my go-to person for queries in the application of Deep Learning to Computer Vision. Vivek has a strong affinity towards research even though he equally values applied research in the context of start-ups and building products in general. We had common ideologies when it came to AI research which helped me strike a strong relationship with him. Even though my internship at Artifacia has ended, I am regularly in touch with him. We frequently share ideas on solutions to many interesting and hard problems in AI.

My relationship with Artifacia

Artifacia's ML team knows its stuff at great depth. I had the privilege of working with enthusiastic bright minds that helped me when I got stuck with issues ranging from understanding existing solutions to implementing them. Working with such a strong team of engineers and innovators helped me improve myself on a great scale. My quantum of learning reached its first peak when I was at Artifacia. Before Artifacia, I was hesitant about my skills to implement deep learning solutions from scratch. Working at Artifacia taught me how to do this and cleared my self-doubt.

My relationship with Artifacia is continuing even after the internship ended. Founders of Artifacia have continued to help me connect with some great young minds across the globe working on AI. I owe a huge portion of my current network to Artifacia which helped me build and expand it. I'm currently working on some really interesting problems in Deep Learning as a Research Assistant at Brown University and I have no doubt about the role my rich internship experience played in my selection for this program. I hope for this relationship to be a long-lasting one. The winter internship at Artifacia gave me some of the most productive and amusing months and I believe this was an experience I'm going to cherish for a long time.