Anjali Samani on the evolution of the data scientist role, the rise of ML engineers and MLOps, and how to lead in data science.
Subscribe: Apple • Android • Spotify • Stitcher • Google • AntennaPod • RSS.
This week’s guest is Anjali Samani, Director of Data Science and Data Intelligence at SalesForce. We first met during the early days of Faculty1, one of the leading data science and AI startups in Europe. Anjali helped design and lead the early Fellowship programs at Faculty (these are intensive bootcamps that take STEM graduates and postdocs and turn them into industrial data scientists).
The focus of our discussion is the evolution of the data science role, the rise of ML engineers and MLOps, and how Anjali leads data and data science teams today. This is the time of the year when people traditionally cast around for new job opportunities, so we take stock of the data science profession, and current expectations of hiring managers.
Anjali Samani:
Bootcamps will still give them a lot of the foundational pieces. But I don’t think that that in and by itself, is enough anymore. In part because the competition is so fierce right now. A lot of bootcamps and universities are turning out data scientists at a phenomenal rate. So there, there are just a lot more people available. Certainly for a lot of larger organizations, like Salesforce, and even actually for a lot of the startups, gone are the days when people want to invest in separate data engineers, machine learning engineers, data scientists. Especially if they can find – they’re not so rare now – unicorns with both data science and some of the ML engineering skills. Organizations are sort of increasingly raising the bar in terms of what they’re looking for. So that’s why I think that just doing a bootcamp in and by itself is no longer sufficient.
Highlights in the video version:
- Introduction to Anjali Samani, Director of Data Science and Data Intelligence at Salesforce
PhDs, boot camps, and industrial data scientists
Boot camps, real world data, and real world projects
What separates a great data scientist from not so great ones?
Any boot camp recommendations for a PhD who wants to be a data scientist?
Boot camps help build your personal network
How would you design a boot camp today?
Finance industry, monitoring, and feedback loop
Responsible AI and risks involving models
Pros and cons of working for a vendor, consulting, or a deep data science role
Are there specific problems your team is working on using bleeding edge techniques?
Time-series and forecasting
What makes business time series forecasting challenging?
Forecasting and optimization
Reinforcement learning, time series forecasting, recommenders, and leadership roles
How would you build a data team from the ground up?
Tools for building models are much more accessible
Do you discuss Responsible AI topics at your current role?
Optimization, experimentation, and basic stats
Making optimization accessible through modern data platforms and opportunities
Related content:
- A video version of this conversation is available on our YouTube channel.
- “Top Places to Work for Data Scientists”
- “Top Places to Work for Data Engineers”
- Jike Chong and Yue Cathy Chang in conversation with Jenn Webb and Ben Lorica: “How To Lead In Data Science”
- Sean Taylor in conversation with Jenn Webb and Ben Lorica: “Changes to the data science role and to data science tools”
- Rumman Chowdury: “The State of Responsible AI”
FREE Report:
[1] Ben Lorica is an advisor to Faculty and other startups.
[Photo by PIXNIO.]