Name: Anuj Diwan
Company: Adobe Research India
Location: Work-from-home (originally Bangalore)
Work Profile: Machine Learning Research Intern
Selection procedure: Adobe India offers two kinds of positions for its internship program: Research and Product. If you have an impressive resume with a high CPI, you can get a direct invite, and I was selected this way for the Research internship. Other applicants were selected via a coding test followed by an interview.
Why Adobe?: I am deeply interested in research, especially in AI and machine learning. After my second-year university research internship as well as a few research projects here at IIT Bombay, both of which fall under university research, I wanted to explore how research happens in a company research lab, to get a holistic view. Adobe is one of the few companies offering an internship at IIT-B in machine learning with reasonable freedom in choosing a problem statement and its solution, which was something I valued. I had also read very good reviews about Adobe’s strong global research internship program and that many mainstream Adobe products used by people all around the world first started out as prototypes developed by their RnD labs, which was really exciting. These two put together, applying to Adobe was a no-brainer.
Work description: Being a large organization, different intern teams get assigned different broad areas of research that Adobe is interested in (to improve their existing products as well as expand their repertoire). I would not be able to write much about my own work due to a Non-Disclosure Agreement, but my problem broadly involves designing algorithms to train machine learning jobs cheaper on the cloud. The first 2 weeks are spent in reading prior work, drawing up a strong problem statement, and defending it in front of the entire research team! This was a great experience that improved my soft skills. After this, a few weeks are spent in devising and implementing a viable solution, usually incorporating iterative feedback from Adobe’s business units as well as other Adobe researchers, finally culminating in a live demo and presentation. Many teams go on to publish their work at prestigious conferences or file patents.
Experience: This internship was truly a unique experience given the pandemic and work-from-home, but as exciting and fulfilling as I had originally imagined! The average workday consisted of reading research papers and articles to get new ideas, discussing them with the team, and coding them up. Although communication and networking within the team were much less frequent than what it would be in a work-from-office internship due to the video conferencing ‘barrier’, we made it work as well as we could, with each team member actively participating in discussions to keep the conversation flowing. Adobe also tried to make up for the inevitable lack of networking with other interns and other Adobe researchers by organizing many activities: quizzes, tombola, informal conversations with the CEO (yes, the CEO himself!) and other leaders of the company, wellbeing sessions discussing stress, work-life balance in a work-from-home setup, and more!
Finally: In any internship, I believe that the experiences gained and new skills learned are the most useful takeaways. In this internship, I learned many valuable things: how research problems that benefit real users are formulated, how to work effectively in a completely new environment (remotely!), and of course the technical and academic knowledge that I gained while doing the internship. While the pandemic may have caused quite an unexpected twist in how the internship played out, I believe that great experience, improved skills, and engaging work reasons enough for satisfaction.
Department Academic Mentorship Programme 