How a professor accidentally started making a Jarvis for himself (and others)

A short tale of the RAx dream

Summer of 2013. Two brothers were on a road trip to St. Augustine from Gainesville (Florida, USA). One of them had recently settled in as a professor in one of the tier-1 tech universities in India. Returning back to India was never easy, and among several odds something struck him upfront — Indian academic research is plagued with three major diseases:

1. It promotes a peculiarly intense yet covert form of exclusivism and elitism and thereby, resulting in power-clique-driven research.
2. It is extremely bureaucratic, where power-politics dominates and influences the spirit of research.
3. It officially incorporates a brownie-point schema that promotes quantity over quality, while desperately trying to catch up with the world-wide “publish-or-perish” rat race.

Not to say that other national academic systems are completely bereft of these diseases, but India surely is in the top charts. This was so much disturbing to this young professor, fresh out of his PhD, that he decided searching for a non-judgmental and open platform that could connect him to talented and enthusiastic young researchers across the country. Unfortunately, he could not find anything that could be of help. As his brother drove past the orange orchards and gator farms of Florida, this newbie prof kept thinking how things can be changed …

India — a new beginning (or madness??)

August 2013. The professor embarked on an adventure which he would eventually be calling Rygbee Campus. The company, Rygbee Inc., was formed quite later in January 2015. Friends keep asking us the meaning, so here it goes — Rig: to drill oil (i.e. knowledge) from ocean; Bee: doing it collaboratively like bees. The idea was to build a research paper discovery platform based on the ongoing research activity of any researcher. In other words, a research stage aware Google Scholar. This would in turn open up offline collaboration opportunities where researchers can find the most suitable potential research students, as well as peer collaborators, for their ongoing projects. A bunch of very talented undergrad students teamed up with him and started the Rygbee movement in the campus of Dhirubhai Ambani Institute of Information & Communication Technology .

The idea was to build a research paper discovery platform based on the ongoing research activity of any researcher. In other words, a research stage aware Google Scholar [for open collaboration].

There was no funding to support anyone. It was pure volunteer work out of sheer passion and belief in the dream. The team got a small store room for themselves (thanks to the university) to work from and they worked real hard — sweaty in the Indian heat without an AC. But this dream would not stay for long …

First step towards disaster: Wrong assumptions

Assumption 1: Lot of researchers want open collaboration but they do not know how to.

What the team found out around June 2015 (we were quite late actually) was that, by and large:

1. Researchers do not collaborate with strangers (and if at all, not for long).
2. Researchers love to work in cliques.
3. Most collaborations are people-centric and not problem-centric.
4. Many researchers are not willing to participate in the sharing process.

Conclusion: We wanted to see a world that the world does not want to be!

A pivot and the second step towards disaster — Rygbee Idea Guide

It was around May, 2016 when the university eCell approached us and convinced us that if we could come up with something that made more “sense” then they would give us some seed funding. No wonder, with a waning team of undergrads busy with their coursework, we started a quest for sense. The desperation of searching sense often results in decisions that are utterly non-sensical, at least from a product point-of-view no matter how magnanimous the goal might seem to be. The leftover team (actually just three of us) came up with this idea of an “Idea-Guide” — a product for everyone who wants to be more creative but struggling for those “aha moments”.

Users would scribble down a bit of their ideas and the platform would “rig up” resources that would trigger further imagination and exploration. Serendipity would be the core value proposition of this product. The “only” thing to be done is to engineer serendipity (sounds like an oxymoron? well, there’s enough research going on in these lines) into the user’s busy schedule.

We formed a separate company called CoRygbee Pvt. Ltd (yeah, we could not get ourselves detached from the name) and got funded (the eCell thought this actually made a lot of sense). But this was a perfect example of a prudent nod from the funding body that legitimised nothing. We still had a very unrealistic assumption:

Assumption 2. Lot of people want to be creative but they need their own pace.

Again, what we found out around May 2016 (really late)

1. Creativity is germinating “shower ideas” usually — people are not serious about what they think are great ideas to be pursued.
2. Creativity will always be a rarity (unless some dramatic genetic mutation happens in the homo sapiens DNA)
3. Thought-journaling behaviour is not a very common habit

Conclusion: We started making vitamin (for everyone) instead of a pain-killer

Introspection time: Going back to what we do best — research

July 2017. We (me and Ankush, my co-founder) had two options left with us — either to get back to our earlier professional pursuits or to figure out a beautiful balance between how we want to see the world and what the world wants to be. What that would mean for us is to build a product for people who have to be creative as a professional requirement. Immediately we felt that it was our own community — academic researchers, and we should be starting from here. If we choose the latter then we have to start afresh. We had no team, no product, no money. We decided we would live off the last penny and make a final attempt. More than anything else, we wanted to give a strong purpose to what we do at a professional level.

The question that was staring at us at that time was “Can we truly help researchers to become lot more productive in their creative process, and not selling our souls into any sort of marketing gimmick?”

But hold on, this presupposes that something is not going great right now with the various ways and tools that researchers use. We surely did not want to build a product that’s just a twist to the familiar tale. That would be such a waste of time for everyone! But as we brainstormed over weeks we realized that the untold pain of the various trials and tribunals of academic research is not unknown, but rather widely unaddressed by all existing tools. The focus so far has been too much (and for long) on either document organisation or citation management. I have earlier written an article in The Academic Rollercoaster about my personal experience with all such tools. Hardly anybody has bothered to recognise that the real demons that slow down research are all at the cognitive process of doing research — they have very little to do with systematic document storage facility and citation styling. I see there are four battles that researchers keep fighting every day:

1. The biggest battle is against our intrinsic limitation of memory and thereby, the difficulty of connecting the dots at a deeper level easily.
2. The second major battle is to give structure to our ideas, observations, conclusions, etc. on-the-fly in a uniform customisable format that is easy to comprehend for any other researcher, including ourselves.
3. The third battle is to quickly figure out what should be the next best read/watch based on our current situation. Situation can be based on whether the need is to explore or to dig deeper or to understand. Situation can also be based on whether the researcher is at the problem definition stage or approach designing stage or evaluation stage.
4. The fourth, and by no way the final, battle is to seamlessly integrate the researcher’s online activity with that of her offline activity (which actually is always the larger part). Ideas do not just come while reading papers online but also during discussions in lab meetings, conferences, cafes, and dreams too!

The birth of RAx — a Jarvis in the making

It wasn’t too long to realize that we have embarked upon a journey of building a Jarvis for academic researchers. We had to scrap off everything, get back to the drawing board, and start all over one more time.

We sensed it early that this is going to be a very different beast as compared to the earlier Idea Guide. The need is so different, so specific, and so intense. In September 2017, RAx was born (RA as in Research Assistant).

RAx is AI powered in the sense that it dynamically senses and adjusts itself to the various subjective needs of the researcher as she moves across the different stages of her research lifecycle — from problem finding to hypothesis formulation to method designing and experimental setup. The current (baby) version of RAx does this in four ways:

1. It helps the researcher in discovering the right research articles for review based on her current research focus and preferences.
2. It helps her to understand the literature she is currently reviewing through supplementary resources such as research blogs, tutorials, video-lectures, and wiki articles.
3. It helps her to ask the right kind of research questions and provoke her to systematically organise the corresponding artefacts scattered across the paper she is currently reviewing, by providing her with customisable review template that follow conventional research methodology.
4. It helps her to easily connect her newly developed insights with your previous reviews and current work-in-progress by providing her with customisable draft templates along with the papers under review.

But more than that, it is getting designed in a way that adheres to how researchers actually work. After all, researchers do not need a new “disruption” in their already disrupted lifestyle! :-)

We are aware that we have to bring in extreme design ideas at the design as well as research fronts. Also, we are fully aware that the baby RAx has the risk of bearing some resemblance of a reference manager if we fail to convey its use cases properly. To that need we will be starting a series of articles in RAx News focusing on specific aspects of RAx and why it has been designed in a particular way so as to fight off the devils that I mentioned before. Till then awaiting your comments.