Oct 2013 – Present · 12 yrs 6 mos · Mumbai Area, India

• At Arya.ai, we are building the responsible ‘AI’ cloud for Financial institutions (Banks, Insurers and Financial Services) to expedite the rollout of complex 'AI' that is explainable, auditable and safe to scale. We are solving the critical ‘AI’ acceptability and scalability problem for FSIs.
• To solve this, we divided the platform into three sub-products -
• Arya APIs - represent the collection of pre-trained models required by BFSIs at one place.
• Libra - provides a repository of task-specific SOTA deep learning models that can be fine-tuned on users training data.
• AryaXAI - the ML Observability tool that plugs with any MLOps to deliver explainability, audibility and safety of ML models for mission-critical 'AI's.
• Today, we work with leading BFSIs in India and SEA. We process the highest number of customer onboarding APIs in GIs in India. We also grew substantially last few months.
• With our new product, AryaXAI, we focus to democratize responsible and safe AI across techniques.
• Arya.ai’s founders are named as ‘Forbes 30 Under 30’. Arya.ai was also named as ‘Top 61 global companies’ by CB Insights and received ‘AI Game Changer’ award from Nasscom. Arya.ai was part of leading industry accelerators organized by Swiss Re, SOCGEN, Microsoft & Oracle. Arya.ai was one of the initial members of Intel AI Builders program and also Nvidia Inception program.

Apr 2013 – May 2015 · 2 yrs 1 mo · Mumbai Area, India

• In academic research, the first step is ‘Literature Review’. Researchers would spend 3-9 months to read papers go through the developments and shortlist a topic for their research. In 2013, more than 30 peer reviewed papers published every minute and Google Scholar and Scopus were simply doing a keyword. Knowledge overload is only going to get worse.
• ‘Knowledge’ suffocation is real and poses serious challenges to ‘Expert skills’. So, we started InvenZone in 2013 to solve the ‘search’ problem and to provide a smart ‘AI’ as an research assistant to STEM researchers.
• With a team of 7 people and Seed round in 2014, we built multiple Language models using Deep Learning during this time and built an ‘AI research assistant’ which can answer complex questions and very powerful search engine that can not do contextual search but also provide a very smart ranking of papers by predicting which research topic would go prominent in coming 3yrs and 5 yrs. We scraped the internet and indexed more than 30 million papers (almost 80% of database at that time) to build these language models. We tried multiple new ways to leverage this data.
• We pivoted from that idea as it had commercial viability issues and users were expectating an ‘Assistant’ that can provide a free-flow conversational debates. DL was not matured to deliver it at that time (2014/15) and capital was not available to deliver this idea completely .

Dec 2011 – Apr 2013 · 1 yr 4 mos · Mumbai Area, India

• Spaneda was one of the first Ed-Tech platforms for students and professionals. I was one of the early employees of the startup, working on coursework, digital marketing, content and student affliates from colleges. I was in my 2nd & 3rd year of college while working in the Startup. It was my professional setup and I really loved working with the team and on the problem, which at that time was very new

Jul 2011 – Mar 2012 · 8 mos

• I wrote this novel in my 2nd year of college when I was 20. It was published by 'Vitasta Publishing', one of the large publishing houses in India and marketing and distribution rights for the novel taken up by , ‘Time Group Books’ a division of “Bennet, Coleman and Company Ltd”.

May 2009 – Jan 2010 · 8 mos · United States

• This was my first book. I wrote this when I entered IIT Bombay in my first year. It was published 'The Ark and the Aroma of Peril' in 2010 with Createspace, Amazon.
• It was not a great one thou :)