Bio

Consider this my business page.
This page will continue to evolve.
Do note the last updation date at the bottom.

PROLOGUE

I consider myself an ambivert geek. Hyper curious, and frequently distracted. I currently handle a data sciences vertical at CoinTribe, in early days of my role I used to make models everyday to predict risk; recently I handle the product pipeline for all things analytics.

I am a cross disciplinary person, juggling leadership, business, product, development, client and research meetings on any typical day. I prefer to sit back and make models in my own time, however, I have developed ways to enjoy and use meetings to be more productive in last few years.

I like to be selective in working only with great bosses, typically who have 10x more achievements and experience than I. I also believe in cutting short on work relationships that are not harmonious. All gains in life are from compound interest, short term thinking is a waste of everyone’s time. Better now work with anyone who you don’t like sitting next to.

Regarding future of work – I acknowledge that an individual cannot be a mere ‘manager’. They must be able to deliver products in by themselves. Shipping beats perfection and Intellectual rigor beats like-mindedness in small teams. I dislike strategy and requirement creation roles, either you are hands on or you are not. Its better to be former for as long as possible. I believe in this deeply and pursue a practice of churning out MVP with least amount of external help.

I respect that organizational goals are considered sacrosanct, however learning should always be aligned with the individual roles. Sometimes, this can be sacrificed in short run to ensure survival, but this cannot be the status quo. The individual should always be on the path of learning and to get better at what they are interested. Otherwise, sooner than they know, they will be on an dead-end path, cursing their bosses, become a mis-fit, throwing excuses to coming in late and generally cutting corners. This flows over to their private lives as well.

Values are everything. The stronger your values, more I am attracted to you. It is not just about lip service, or display of corporate slides with ‘Core Values’ CEO shove in employees faces. Values evolve and should be battle tested. People will genuinely put more skin in the game if they believe in your values. In my experience this rates far above any monetary reward for loyalty. If we want to work together, its better to get understanding of our values and how far they are aligned. More trust, more transparency, better payoffs. It is just better business.

Finally, I have this core belief that everything is an algorithm. Any finished product is an algorithm. An algorithm is the fruit of discovery. In Machine Learning an algorithm is also referred as a Model.

It starts out with a story, usually in form of a question or a problem looking to be solved. Becomes an arc with gathering of characters(data points) and mutual exchange between these data points tells us which is the hero and who’s the villan(in the context of the story). This give way to our finished algo. Connecting the inputs to an satisfactory and easily digestible output. In my experience of data science, domain matters little, process matters far more, mathematical fundamentals matter the most.


FAQs about my work credentials:

  • What you are Primarily?
    • A Machine Learning Practitioner
  • What roles you have filled?
    • Data Scientist, Analytics leader, Data Science development Head, Data Consultant to CEOs
  • What can you do for me?
    • Handle your Data Science vertical, Make an killer algo with your existing data, Discover patterns stored within dusty data stored on your private cloud, Predict risk relevant to your business model.
  • Can you code?
    • Of course. Code is speech in land of software.
  • In which language?
    • Any (given 2 week headsup). Primarily, Python and R.
  • Do you know #AI?
    • Yes. It is just an emergent branch of ML. Don’t read the hype, 90% of AI are just regression models.
  • How about #DeepLearning?
    • Sure. But the results would be hard to explain. It just works(magic and hidden layers)
  • What are your Go-to Algorithms?
    • Gradient Descent and PCA
  • Can you travel for work?
    • Sure but temporarily
  • Do you consult?
    • Yeah
  • Would you consult for free?
    • Sometimes (just ask)
  • Can you help setup entire data science vertical?
    • Give me 2 months
  • What projects have you worked on?
    • Risk, Recommendations, Anomalies, Time series, Stock price prediction, Demand Forecasting, Image recognition
  • You mentioned DeepLearning, what projects?
    • Handwriting recognition, Object detection, Image to digital text extraction.
  • Where you see yourself in 5 years?
    • Making movies
  • Ok. But Seriously..
    • Yeah seriously, make movies, or make stories for them.
  • What do you fear?
    • Stasis
  • Are you a happy person?
    • I am a rational optimist
  • What companies you have worked for previously?
    • CoinTribe, Snapdeal, Goibibo, MachineParty, PTC
  • Why no big companies?
    • I prefer small and fast paced teams
  • How much total experience?
    • Just over 6 years
  • Any experience on giving talks?
    • Gave talk in Dubai in 2016 on industrial applications of ML, Keynoted an #AI conference in Hyderabad in 2017, Spent some time doing research in Germany and France (mostly Computational Physics)
  • What college you are from?
    • IIT-Delhi
  • What Major?
    • Engineering Physics. B tech.
  • Where did you get the rigor for Machine Learning?
    • Been always interested in maths and probabilities, my major at IITD taught me a lot of fundamental maths, my internships in Germany and France added the applications to my arsenal
  • Any interviews you gave?
  • How can I contact you?
    • skype: rahul.kapil or mail given in footer address
  • Do you want to do startup together?
    • No
  • What are your hourly consulting rates?
    • Depends on task (from free to 200$)
  • What else you are interested in?
    • Storytelling, evolution, astrophysics, mental models, football, TT, compression, information theory, teaching.
  • Do you sing or dance?
    • No. But I secretly wish I did.

 

updated: 27/11/2017