The deep leap

While everyone in IT industry is talking about getting ready for the future, reskill yourself, I finally took the courage to do it last year.  In September last year, I took a decision to let go of comfortable job I had been in for more than 10 years and do something I had been looking at from the edges.

---

I always considered Machine learning as one of the most promising areas, one that will change the future of mankind and I usually consider it at par with inventions like the wheel, paper, internet, and mobile. But I was missing to be part of this future.

Given my inclination and interest, I left my full-time job 4 months back and took the leap to learn and implement machine learning.

I joined a lesser know startup Mindbowser and got a chance to build the Machine Learning practice from scratch. Chose Mindbowser out of 4 startups I had been speaking with because Mindbowser seemed most keen to try out new things and go in for the long run.

Initially, I started off with learning about ML on my own and getting subscribed to the online Stanford course (https://www.coursera.org/learn/machine-learning). The course was pretty good in terms of content for understanding basics but without the real world scenarios, the theory tends to become a misty.

So what we did was to knock a few doors and ask if they would like to check out how ML can help their product for free. Convincing companies to give their data was hard but that was the lesson learned- You have to be a hustler. Have to build a story when none exists and sell it to the world.

I connected back to my old contacts and startups to ask them for work and something clicked.

In November we had our hand on our first task- An IOT startup had some set of sensor units that were sending signals frequently. Our task was to receive the signals & parse the signal data. Based on the data from each unit we need to find out outliers from the data. For the Novelty and Outlier Detection, we used the opensource sci-kit-learn libraries and were able to provide accurate results up to 95%.

In the meantime, we also received an interest from an existing Mindbowser client, a market research company that wanted to generate some statistics for the product, based on the inputs in the survey. This task helps me to improve my statistical skills.

A very interesting task that we recently picked up is to predict the sentimental analysis for a healthcare agency based on their social media activities. This task is divided into various segments like Web Scrapping, Text Analysis, Image analysis etc. I am feeling very enthusiastic about this task.

As you would easily say these were some really good opportunities I got to work with within a very small time, things that probably I would have missed I would have waited for someone to reskill me.

All this while I also got to learn some great life skills such as the importance of network and the way to hustle for opportunities. From leaving a cushion job at one of the large IT companies today I am able to invest time in myself and things I want to do in the next coming years. I am discovering myself to be more disciplined and innovative than I thought I was. I also got to learn about learning from others and the importance of mentors. I have become open to helping others and be a mentor as well. In fact, I am mentoring a full new crop of ML now.

In a talk given by Jack Ma, he insisted that being best is not the most important thing. Being there first is. We are living in a world full of opportunities and challenges. Can we try to be the first to solve one of them? I think now I understand what he meant.