Shivaram K R is currently Co-Founder and CEO of Curl Analytics. He has an academic background. Shivaram is a Machine Learning (ML) expert. He has been building ML models from last 12 years. He started mastering ML in IIT Delhi. Since then he has worked on various kinds of problems in multiple domains. Shivaram has a keen interest in Maths, development of new algorithms and solving challenging problems. Before starting Curl, he was working in the Investment Banking (IB) domain for 10 years. He has worked as a Quantitative Analyst and Algorithmic Trader. Throughout his career, Shivaram’s job was to forecast and estimate the financial markets. He has worked in IB in some of the world-renowned companies such as Bank of America, Edelweiss Securities and Societe Generale.
How did you get into Data Analytics? What interested you in learning Data Analytics?
Shivaram K R: I did my M.Tech in Control and Automation at IIT Delhi. Control and Automation deals with controlling systems using various techniques. I choose that particular specialization because of Dr Madan Gopal (an authority in that field) who was teaching there and I liked robotics control. Dr Gopal had started using ML to control various systems instead of conventional techniques. ML and its power to learn the problem using the data fascinated me. I took many courses in ML and Computer Vision. I also took classes on embedded electronics in my spare time. Professors in my department liked my work and gave me keys to one of the labs. My main project was on text data analytics using SVMs and other feature selection methodologies.
What was the first data set you remember working with? What did you do with it?
Shivaram K R: When I was studying in IIT Delhi, there was an open house event (in 2006) where people from outside could come and see our work. I was given the responsibility to make it successful for our group. I worked on many experimental setups such as robotic arm, twin rotor etc. I felt I had to make it interesting and interactive. I wanted to use ML and show its power in real time. I got an idea to build a classifier to distinguish male and female voice. I collected audio samples from boys hostel by making them utter specific words. For female voice samples, I remember that I had to bribe girls with chocolates to collect their samples. I built the model and it had decent accuracy. Our open house was a big success!
Was there a specific “aha” moment when you realized the power of data?
Shivaram K R: There are many “aha” moments when it comes to ML. Most recent one was because of Deep Learning (DL) based Reinforcement Learning (RL) model. I play a game called DOTA 2 when I get time (now it is close to impossible). DOTA 2 is a strategy game with lot of complexity. It is more like real world as the number of things one can do in almost infinite. Recently, OpenAI built a 1v1 AI bot, which could defeat the world’s best DOTA 2 players such as Sumail, Dendi etc. This was a shock and “aha” moment for me. It took me two days to accept the fact that bots can be better than human in DOTA 2!
What is your typical day-in-a-life in your current job? Where do you spend most of your time?
Shivaram K R: I am CEO of a start-up that specializes in emerging technologies such as ML, IoT and Blockchain. It is obvious that I have a busy schedule. Adding to that, Curl works on various industry verticals such as finance, IIoT and genomics. We are growing very fast and I have many responsibilities. My typical day at office is something like this:
- Talk to other Co-founders during the morning walk
- Write Quora answers when I have breakfast (I am currently 2nd most viewed writer in Machine Learning in the world)
- Go to office and spend time with algorithmic strategy team, video analytics team and IoT team. I typically ask them to show their progress and give ideas to solve their current problems
- Talk to the CFO to figure out how we can expand and hire more people
- Create strategy for executing various projects in hand. Work on other possible leads
- Build algorithmic trading strategies or work on algorithms for analysing genomics data
- Read technology related news and advancements happening in ML and DL
- Take breaks to teach Curlites on various topics such as computer vision, trading etc
- If some project is running late, I sit at their desk and start coding
- Inspiring Curlites to go beyond their capacity to achieve absolute perfection. I call it becoming a true artist; it is a level above experts.
- End of the day I hit the gym and I come back home to my lovely family
When I am not in the office, I am meeting clients to get business, building contacts or giving talks on ML at various institutions to inspire students. I give talks to spread the word that they can join Curl if they want to do high-end research in India. I spend almost all my time awake thinking about Curl and how to take it to the next level.
How do you stay updated on the latest trends in Data Analytics? Which are the Data Analytics resources (i.e. blogs/websites/apps) you visit regularly?
Shivaram K R: I like reading about other science-related topics, technology and markets. Here are few of the links:
I like reading latest academic papers. Here are few links:
- Two-minute paper:
Share the names of 3 people that you follow in the field of Data Science.
Shivaram K R:
- Yoshua Bengio
- Andrew Ng
- Geoffrey Hinton
Team, Skills and Tools
Which are your favourite Data Analytics Tools that you use to perform in your job, and what are the other tools used widely in your team?
Shivaram K R: We use Python for building our models. We build models from scratch. We use ML and DL libraries such as Tensorflow, Pytorch and Scikit-learn. For Blockchain based projects, developers use Etherium, Corda and other platforms
What are the different roles and skills within your data team?
Shivaram K R: We have expertise in ML and DL, financial markets, IoT data, Blockchain, cloud and big data. The roles are video analytics, IoT data analytics, Quantitative analyst and R&D specialist.
Help describe some examples of the kind of problems your team is solving in this year?
Shivaram K R:
- We are building an algorithmic strategy platform and developing various strategies for some of the biggest funds in India
- We have developed a new pipeline for HLA typing using data analytics and are publishing a paper
- We are working with Karnataka government to forecast rainfall over 3 month period
- We are working with Karnataka and Andhra Pradesh government to analyze traffic and road quality.
- Helping AP Government in their e-governance. Publishing a paper on Telugu language analytics.
- We are working with NTU (Government of Singapore) to develop end to end IoT solution with edge computing and ML-based analytics
- Worked with various industries in their automation using vision based solutions
- Working on IoT and Blockchain based solution for ports
How do you measure the performance of your team?
Shivaram K R: I talk to Curlites on a daily basis. I understand their problems and help them wherever needed. We hire talented people with right attitude. Most of them don’t need any performance evaluation as they generally exceed expectation!
Advice to Aspiring Data Scientists
According to you, what are the top skills, both technical and soft-skills that are needed for Data Analysts and Data Scientists?
Shivaram K R: ML and DL need three major skills. Maths, programming and domain knowledge. Maths and Programming are common to all the problems. One need to read lot of books, attend online courses, read blogs, participate in hackathons, read latest papers and implement them to master Data Science.
Beginners start with online courses and books. You can read my Quora answers for more details.
How much focus should aspiring data practitioners do in working with messy, noisy data? What are the other areas that they must build their expertise in?
Shivaram K R: The real world data in very noisy and it is important to understand that. It is good to start with Kaggle problems and work your way from simple once to difficult once. You can start participating in hackathons and code challenges. Math is very important to understand and fine-tune the models. Making the habit of reading papers is very important.
What is your advice for newbies, Data Science students or practitioners who are looking at building a career in Data Analytics industry?
Shivaram K R:
- Programming and software skills: I suggest beginners to start with Python as it can be used even at advanced levels. You can also explore Julia.
- Visualization Tools: Matplotlib, plotly, seaborn are good python visualization libraries
- Statistical foundation and applied knowledge: Read math section in books such as Elements of Statistical Learning and Pattern Recognition and Machine Learning
- Machine Learning: Go through courses to learn
What are the changing trends that you foresee in the field of Data Science and what do you recommend the current crop of data analysts do to keep pace?
Shivaram K R: The field of data science has moved into Deep Learning (DL). The DL models have state of the art accuracies and are really changing the world. The new data scientist should work on deep learning models such as CNN, DNN, LSTM, Memory models, GAN, RL etc. To keep pace, read latest papers in arxiv.org and follow people who are working on DL.
Would you like to share few words about the work we are doing at Digital Vidya in developing Data Analytics Talent for the industry?
Shivaram K R:
We are working closely with Digital Vidya. We like their approach of hands on practice and capstone projects. We are giving industry insight and helping them create a syllabus, which is future proof. We have plans to mentor students and take some students for internship at Curl.
Are you inspired by the opportunity of Data Analytics? Start your journey by attending our upcoming orientation session on Data Analytics for Career & Business Growth. It’s online and Free :).