Careers in Machine Learning and Data Science | Q&A with Vin Vashishta

Machine Learning and Data Science are making a huge impact all around us. However, they are also overhyped. Another issue with the area is young folks are not always sure about data science or machine learning career path. In this article, Vin (Vineet) Vashishta, one of the top Machine Learning and Data Science experts in the world, will talk about careers in machine learning and data science. He was one of the LinkedIn Top Voices in 2019 and also among the top ML influencers on Twitter.

Machine Learning Engineer and ML Product Strategist

Vin is a Physics major. He builds custom machine learning models with a focus on productization. His current venture (V2) started out as an extension of his personal consulting practice.

Now the business is product-focused. They have two product lines, several datasets, and data acquisition services.

Here are a few candid and brutally honest career advice for aspiring machine learning engineers and data scientists.

Careers in Machine Learning and Data Science

Q&A with Vin Vashishta

 

Tanmoy: You are a Physics (major) graduate. You opted for an MBA after a decade of work-ex.
At present, you are a core data scientist, machine learning developer, and consultant. How would you describe your career journey so far?

Vin: I see two paths to value in our field: Applied Research and Applied Development. I have built for production my entire career so that is where I landed in the field.

I saw technology adoption fail from my earliest days in tech in the mid-’90s. I saw products fail to gain traction in the market, good products. I saw companies fall because their business model was disrupted by companies that could rapidly adapt and apply new technologies.

The strategy side is the barrier for most businesses. I have been a bridge between the C Suite and the technology. Sometimes that meant hands-on building to prove the value of machine learning. Sometimes that meant building teams. Sometimes that meant building a culture and buy-in for adoption.

My contributions to the field are getting companies to spend money on machine learning, defining the applied, business-facing role, and making the field more accessible.

My time in this field is coming to an end. I will be moving on to the next technical advance with the same objectives. My experience with pushing machine learning forward in business will help me be more effective with my next phase.

 

Tanmoy: There are many seasoned professionals who feel the need for upgrading and reskilling through graduate programs. How should they approach such a situation?

Vin: Learn, apply, mentor. That is the new capabilities reskilling cycle.

College is a failed construct. Abandon it.

Find businesses that foster mentored, project-based reskilling. They should have a roadmap for capabilities advancement that aligns with business goals over the next 3-5 years.

Companies that fail to build reskilling models will not be in business much longer. Do not bother working with them except to achieve short-term career goals.

Related Article: Importance of Reskilling and Upskilling

 

Tanmoy: How exactly does V Squared (V2) help firms with their businesses?

Vin: Applied machine learning. That takes a full machine learning lifecycle. The business case to production and maintenance. I work with companies to advance their ML workflow and lifecycles.

Machine learning maturity. Moving from prototypes/proof of concepts to simple business cases, to high ROI customized solutions to create a competitive advantage. Creating culture, executive-level sponsorship, and accountability for results.

Rebuilding the business model from static competitive advantage to transient competitive advantage. The latter is built to adopt new technologies. Those companies profit from disruptions to the marketplace and opportunities to enter new markets.

 

Tanmoy: What would be your advice for students who want to make a career in machine learning and data science?

Vin: The field is yours. It may seem like there has been a lot done in the last 10 years. In reality, we have put down a foundation and cornerstones. What comes next is up to you. I am working to define a framework for the field and a mid-term (5 years) path forward.

What will define the field over the next 25 years is up to you. One of the reasons I am going to move on is to get out of the way.

Tanmoy: There are many critics who think Data Science, ML/AI & Big Data Analytics are overhyped and are just buzz words.
They think (as per my understanding) those fields are just advanced software programming & computer science, statistics & mathematical modeling, or fancy terms for Business Intelligence & Market Research.
What are your thoughts on this?

 

Vin: I agree

Tanmoy: What are going to be major trends in Data Science, ML/AI and Analytics in the next 2 – 3 years?

Vin: Automation for static competitive business models. Products based on NLP, Computer Vision, and Bioinformatics for transient competitive business models.

A move towards team machine learning; Data Analyst, Data Scientist, and Machine Learning Engineer cross-functional teams. Advanced R&D will add a Machine Learning Scientist.

As businesses move forward in the ML maturity model, ML Product Managers, ML Test Engineers, and ML Maintenance Engineers will be emerging roles.

Customized models will become the drivers of value. Import from… will move to data analysis and visualization support. As businesses move forward in the ML maturity model models will progress from descriptive to predictive to prescriptive. Businesses will expect prescriptive models and see the rest as analytics.

Tanmoy: Due to the hype, many folks think of shifting to Data Science, ML/AI, and Analytics.
Many non-CS grads (Mechanical or Electrical-Electronics grads), who do not have any relevant coursework or practical experience, want to pursue MS Data Science or ML/AI.
What would be your advice?

Vin: The most common paths into a Data Analyst role are from Analyst, Business Analyst, and academic Researcher or Research Assistant roles.

Data Scientists transition from Data Analyst and academic Researcher or Research Assistant roles.

ML Engineers transition from Software Developer, Software and Cloud Architect, and Data Scientist roles.

Machine Learning Scientists transition from academic Researcher or Research Assistant and Data Scientist roles.

Paths into the field involve these capability reskilling progressions. Experience rules the day.

Years of experience do not correlate with employee performance. The measure is an application from the Learn, Apply, Mentor reskilling model.

During the learning phase, there are guided project work. Once that work meets a minimum standard of application, that person has become a Data Scientist.

Similarly, there is no correlation between degree and employee performance. Again, college is a failed construct. Abandon it.

The focus on a degree does not lead to any sort of ROI for a company or the individual.

 

Recent Articles by Vin Vashishta: 

 

Machine Learning for the Return to Work – Reskilling Using ELV Models

What Is The New Normal For Data Science Hiring?

Data Wrangling For Machine Learning Professionals Part II – Data Privacy and Security

 

Tanmoy: How should learners leverage online courses (MOOC platforms)?

Vin: I do not see any of those platforms providing a capabilities-based curriculum. They effectively teach concepts.

Core concepts are important.

Linear algebra or software development fundamentals are important core concepts. Concepts in a vacuum will not get a person hired or provide value to a business.

 

Tanmoy: What are the top technical skillset for a Data Scientist (Python, Tableau, or anything else)?

Vin: Skills are meaningless. What can you build? What can you push to production? What value can your models bring to the business? None of those are tied to a word like Python.

Skills are a result of a disconnect between strategy, execution, and employee capabilities/organizational capabilities. Talent is a commodity. It can appreciate in value with directed investment. Employee lifetime value replaces skills with capabilities. It ties capabilities to business value. It creates a cycle of continuous, directed reskilling.

 

Tanmoy: How can non-technical graduates (Business Studies or Humanities majors) end up in the core Tech sector? Do you have any specific career tips for them?
Secondly, how important are soft skills in the core tech domain?

Vin: Focus on developing and adapting the capabilities you have built and strengthened, not the ones prescribed to you. In an employee lifetime value paradigm, critical thinking, communications, teaching, leadership, time to apply learned capabilities…these are core strengths.

Core tech emphasizes simple technical skills.

Skills like Python or Cloud Architecture are over-glorified. The gatekeeping within tech makes technical skills seem complex but they are some of the most basic. I have met a number of Principles, Distinguished, and Senior Staff level engineers who have made their whole career reusing the same 2000 lines of code and overcomplicating their work to create job security.

There will be a wave of new businesses in the next 2-3 years driven by non-tech founders who focus on solving complex problems with business-centric solutions. They will dominate those companies that have become complacent providing technology-centric solutions.

 

Tanmoy: How important is personal branding in today’s job market and what are your suggestions on this?

Vin: Personal branding is very important. It is one of the most effective ways of differentiating and marketing yourself. A brand creates a pipeline of opportunities.

 

Creating a brand is a lot of work. It requires an individual to create something unique. Being a copy of someone else is not a personal brand. Why buy Nike over Reebok or vice versa? The differentiation is perception-based. They create a following who connect with their core, defining characteristics.

Doing that as an individual means detaching yourself from other brands. Ex-Google is not a personal brand. That makes a person a Google knock-off. Their personal brand does not really exist. A personal brand needs to be individually owned and originating from the individual.

 

Tanmoy: How would you describe your style of mentoring/supervising juniors/employees?

Vin: Here is a project. Come get me when you need help. My office hours are… I will check in with you daily. Struggle but do not keep hitting your head on the same brick wall. You have a deadline.

Bring creativity to your solution. Go beyond my specifications where you feel it is necessary. Disagree with me often. Challenge me often.

Fail and be wrong often. Be accountable for your actions and work products. Just put your hand up and say, “My role in this failure was… Next time I will…to fix that.” I can help you improve your understanding of your role and your steps to improve.

Hold me accountable. If you fail and I have not anticipated it and put a safety net in place, shame on me. How did I not see that coming? I should be as accountable to you as I expect you will be to me. I should be as willing to learn from you as I expect you are from me.

Your capability to apply what I am mentoring you on is a critique of my effectiveness in that role. We are not alone on this boat. We succeed as a team or fall apart as individuals. I cannot make you do it. You have to want this. I have to make it possible.

 

Tanmoy: When you are not working, how do you spend your spare time – any hobbies?

Vin: Usually I give a canned answer but, in truth, I do not have any hobbies.

I find comfort in my work. I feel an absence of purpose when I am not working. It is very unhealthy, and I encourage no one to follow my lead here.

Maybe someone who used to be like me can suggest a hobby to start with.

 

Tanmoy: Finally,  any general piece of career advice for students (high school & college) and fresh graduates?

Vin: Here it is – 

 

Marshawn Lynch: That’s when it just clicked in my mind that if you just run through somebody face, a lot of people aren’t going to be able to take that over and over and over and over and over and over and over and over and over and over and over and over and over and over and over again. They just not gonna want that.

Interviewer: Think there’s a deeper metaphor there?

Marshawn Lynch: Run through a mother…’s face. Then you don’t have to worry about them no more.

Marshawn dedicated his career to a single mantra. He did not run from the truth of his nature.

 

I recognize patterns and explain their significance. That is all I do. Find your mantra.

Vin Vashishta

Related Articles:

The Evolution of Data Science Career Path

Beginners Guide to Data Science, Machine Learning, AI

Translate »
%d