Even though many new pro­gram­ming lan­guages and frame­works have emerged lately, many de­vel­op­ers still need a guide — ei­ther on­line or off­line — to en­hance their cod­ing skills. Janani Ravi, a former Googler and Stan­ford alumna, closely ob­served this be­havio

OpenSource For You - - For U & Me -

QHow did the idea to launch Loony­corn come about? The seed of the idea came in 2013, when my hus­band and I were both work­ing with Google in Sin­ga­pore. We en­coun­tered a video by RSA (the Royal So­ci­ety of Arts) about what mo­ti­vates peo­ple to work. The video has over 10 mil­lion hits on YouTube. We were blown away by how vividly we were able to recall the video even years after we watched it. After that, we started ex­per­i­ment­ing with a video-mak­ing tool called Video­scribe, and then started to build its au­to­mated ver­sion. By 2015, we had built a pre­lim­i­nary ver­sion of a video auto-gen­er­a­tion engine. We filed for a patent for it in the US as well as in In­dia, but then we re­alised that we en­joyed mak­ing con­tent more than build­ing an engine to help oth­ers build con­tent. So we quit our jobs and started work­ing on build­ing niche con­tent full time. By do­ing this on a reg­u­lar ba­sis, our startup mor­phed from a tech startup into a con­tent startup.

QSo what’s dif­fer­ent about Loony­corn when com­pared to the oth­ers in the mar­ket? We are dif­fer­ent in a good way be­cause we are su­per-small with a team of eight mem­bers, have no ex­ter­nal fund­ing and are fo­cused on just sit­ting down in one place for hours on end and do­ing work that we can be proud of. In about 18 months since we started Loony­corn, we have built about 100 cour­ses and en­rolled more than 100,000 stu­dents. We have also made lots of mis­takes, but they were hon­est mis­takes. I think we have largely avoided mak­ing the same mis­take twice.

QHow do you re­search de­vel­oper-cen­tric con­tent? Un­til quite re­cently, we only made con­tent on sub­jects that we al­ready knew very well and had used in real life. But now we have started learn­ing top­ics and tech­nolo­gies, and mak­ing con­tent based on them. For this, we plunge in and learn the ad­vance­ments as sys­tem­at­i­cally and thor­oughly as pos­si­ble. We play around with a lot of ideas, par­tic­u­larly in the area of fi­nan­cial trad­ing; so we try and find use cases for new tech­nolo­gies re­lated to fi­nan­cial trad­ing.

QWhat are the chal­lenges you’ve faced while search­ing for the most ap­pro­pri­ate con­tent for de­vel­op­ers? The big­gest chal­lenge for me is get­ting out of my com­fort zone and learn­ing new tech­nolo­gies. There is not much of a chal­lenge in search­ing for new tech­nolo­gies; there are some amaz­ing tech­nolo­gies to learn from. The chal­lenge is to learn those tech­nolo­gies and be­come adept at them — it is the chal­lenge of over­com­ing our own in­er­tia to master new stuff.

QWhat, ac­cord­ing to you, are the main is­sues the world of com­puter science faces to­day? One big chal­lenge is that the world is be­com­ing a win­ner-takes-all world.

This is not only in tech­nol­ogy — it is hap­pen­ing in ev­ery walk of life. In the US, for in­stance, so­cial mo­bil­ity is far less than it used to be in the past. Those who have money, skills and tal­ent keep get­ting more. The have-nots are get­ting squeezed. I think all this is largely driven by su­per-spe­cial­i­sa­tion and by the speed at which tech­nol­ogy changes. There is no al­ter­na­tive to learn­ing and in­no­vat­ing. If you don't learn, you are no longer rel­e­vant and slowly fade away.

QDo you be­lieve the In­ter­net is mak­ing things eas­ier for 21st-cen­tury coders? Things aren't get­ting eas­ier for most coders. The In­ter­net makes tal­ent very mo­bile; so, now we all have to keep pace with the best in the world to stay rel­e­vant. It is not very dif­fer­ent from what is hap­pen­ing in the other spa­ces such as the restau­rant mar­ket. A decade ago, a restau­rant needed to be good enough or bet­ter than other restau­rants in a one-kilo­me­tre ra­dius to sur­vive. But now, with apps like Zo­mato and Yelp, and taxi ag­gre­ga­tors such as Uber and Ola, cus­tomers can even travel for a dis­tance of 15 kilo­me­tres to go to a great place to eat rather than sit­ting and tast­ing food at a medi­ocre one.

QHow im­por­tant, do you think, is it for de­vel­op­ers to have learn­ing habits to en­rich their ca­reers? Learn­ing habits are get­ting more and more im­por­tant in my opin­ion. As com­puter science pro­fes­sor Cal New­port writes in his book Deep Work, de­vel­op­ers need to be pre­pared to sit down and do ‘deep work’.

It is sat­is­fy­ing and pro­fes­sion­ally very re­ward­ing.

QDo In­dian de­vel­op­ers lag be­hind in skills, when com­pared with their coun­ter­parts in de­vel­oped re­gions like the US? There is a grit gap and not a skills gap be­tween In­dian and Amer­i­can de­vel­op­ers. The best coders here are ob­vi­ously as good as the best any­where. In fact, we have so much to be proud of in tech. But at the same time, we need more coders who are like ac­tive bats­men. The will­ing­ness to stick with a dif­fi­cult prob­lem, sit in one place for hun­dreds of hours, code up thou­sands of lines of code — we are way be­hind in the per­se­ver­ance needed to do all these tasks. This grit prob­lem is the root cause of the skills gap — tech­nol­ogy changes so fast that ev­ery­one with­out ex­cep­tion is go­ing to find their skills be­com­ing ob­so­lete. The prob­lem is that we are miss­ing the grit to learn new tech­nolo­gies, and pre­fer to con­tinue to sur­vive in the mar­ket with older learn­ings.

QYou’ve in­di­cated that fi­nan­cial trad­ing is one of the fo­cus ar­eas of Loony­corn. What are the ma­jor of­fer­ings in the pro­gram­ming world that help coders to en­hance their skills on that front? My hus­band (Vit­thal Srinivasan) has been a quant trader for quite a while in the past, and his take is that ma­chine learn­ing is the way to dis­til good trad­ing sig­nals. These days, you can rely on sys­tems such as Ten­sorFlow to ease the ex­pe­ri­ence. It is no­tably Python that makes sys­tems smarter for fi­nan­cial trad­ing.

QHow does Python help de­vel­op­ers build so­phis­ti­cated fi­nan­cial mod­els? There is a ton of pow­er­ful stuff that de­vel­op­ers can lever­age to build new so­lu­tions for fi­nan­cial trad­ing. Python has Pan­das, Numpy, Scikit and, of course, Ten­sorFlow, for sup­port­ing deep learn­ing and ma­chine learn­ing in tra­di­tional mod­els.

QLastly, what’s the ad­vice you would like to give de­vel­op­ers who are about to step into the cod­ing space? My brief ad­vice for as­pir­ing coders is to learn to sit down and work long hours — hun­dreds of hours on a reg­u­lar ba­sis. They don't have to worry about money or any re­wards at the ini­tial stage. These fruits will fol­low if they are skilled enough. Here, I will use the term kar­mayogi — work for the sake of the work it­self.

Janani Ravi, former Googler and

Stan­ford alumna with her son

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.