Knowledge Scientist Highlight: Belén Sánchez


DataRobot is residence to a treasure trove of Knowledge Scientists. We’ve began a sequence known as  “Knowledge Scientist Highlight” as a technique to meet the individuals behind the know-how and introduce you to our nice group.

Introducing Belén Sánchez. Belén works on accelerating AI adoption in enterprises in america and in Latin America and contributes to the design and growth of AI options for patrons within the retail, training and healthcare industries. 

1. What novel strategies and methods are you at the moment utilizing?

I’m utilizing so many novel strategies and methods which have been integrated into DataRobot that it’s exhausting to present a easy reply.  However let me share a number of helpful and thrilling strategies that I’ve been utilizing currently:

  • Time sequence – new sequence modelers: One of many challenges that you would be able to face when forecasting demand is tips on how to work round merchandise which can be seasonal or are newly launched.  With DataRobot, I’m now capable of predict new sequence with no historical past  that haven’t been seen in coaching with the introduction of a number of methods that assist to maintain the predictions extra outlined with out wild outlier predictions. A few of these methods embody single modeler with a number of estimators and mannequin collections.
  • Bias and Equity: This subject has been on the prime of my thoughts since I began my profession as an information scientist.  Nonetheless, for years I felt that many discussions on this subject weren’t capable of land concrete steerage on tips on how to work round this.  Lastly, with DataRobot I’m able to have a transparent AI bias and equity workflow that helps me acknowledge and repair bias in my fashions.  This workflow contains the identification of protected options in your dataset, the number of an acceptable equity metric, and the technology of insights to determine and perceive the mannequin potential bias. And it seems to be like quickly we are going to even have methods to mitigate bias uncovered in your knowledge by means of the platform!
  • MLOps: This subject has been related for years, however definitely it grew to become much more related as soon as the pandemic hit us.  So it’s fairly thrilling to be working with state-of-the-art know-how and practices that present a scalable and ruled means to deploy and handle ML purposes in manufacturing environments.  Monitoring accuracy over time, knowledge drift and challenger fashions are definitely a few of my favourite practices.

2. If 85% of fashions fail to make it to manufacturing – how do you take care of failure?

I believe the easiest way to take care of failure is to acknowledge it, settle for it and be taught from it as quick as you may.  Because the query factors out, there’s a excessive share of machine studying fashions that fail, and listed here are a few issues I’ve realized from my very own failed fashions:

  • Floor your mannequin on an actual ache or enterprise drawback.
  • When designing your answer, embody the voices and views of the stakeholders and customers that will likely be affected by your mannequin consequence.
  • Iterate and be sure you can clarify how your mannequin works and make predictions.

3. How do you see the function of the information scientist evolving sooner or later? 

I used to be discussing this with one among my colleagues and we each agreed that within the coming years, we are going to see extra firms and organizations profiting from the workforce augmentation and due to this fact investing within the growth of AI based mostly options throughout totally different areas.  Which means there will likely be extra alternatives to use knowledge science throughout companies and industries, and area or business experience will likely be very worthwhile.  On the identical time, knowledge scientists may have extra alternatives to contribute to the R&D or product growth from a ML engineer perspective.

4. Does DataRobot make an information scientist’s job simpler? How?

It definitely eases some elements of your work as an information scientist, however I believe the advantages of DataRobot transcend that. DataRobot makes you extra productive, it accelerates the pace to construct a superb mannequin, it exposes you to novel and sturdy strategies and algorithms, it permits you to deal with extra strategic issues, it gives options that cowl the entire life cycle and administration of a mannequin and final however not least it facilitates collaboration amongst individuals with totally different roles than yours.

5. Do you might have any ardour initiatives? 

Time sequence initiatives have captured my curiosity and fervour over the last yr.  I take pleasure in engaged on time sequence initiatives with my shoppers.  I believe that having the aptitude of forecasting enterprise metrics resembling gross sales, turnover, web site site visitors, and so forth. brings lots of worth to an organization.  But, it’s also one of many areas that wants extra experience. 

I even have a real ardour for contributing to the discount of the gender hole within the AI business.  This yr I used to be capable of lead a DataRobot College program that supplied  scholarships to 60 Latin American ladies residing throughout 11 international locations. They’ve been studying about AI and utilized knowledge science in a seven7 week Spanish coaching program.  This has definitely change into probably the most passable accomplishments in my knowledge science profession to this point. 


See How DataRobot Extends the Capabilities for Knowledgeable Knowledge Scientists

Be taught Extra

In regards to the creator

Belén Sánchez
Belén Sánchez

Knowledge Scientist, DataRobot

Belén works on accelerating AI adoption in enterprises in america and in Latin America. She has contributed to the design and growth of AI options within the retail, training and healthcare industries.

Belén is a frontrunner of the WaiCAMP by DataRobot College Initiative that contributes to the discount of the gender hole within the AI Business in Latin America by means of pragmatic training on AI. She was additionally a part of the AI for Good: Powered by DataRobot program, which companions with non-profit organizations to make use of knowledge to create sustainable and lasting impacts.

Earlier than becoming a member of DataRobot Belén labored as an information scientist and as a global coverage advisor on financial development in organizations just like the World Financial institution, the Inter-American Growth Financial institution and authorities businesses in Latin America.

Belén has a Grasp in Public Administration from Harvard Kennedy College and a M.B.A. from the College of Leipzig in Germany.

I’ve change into a greater knowledge scientist through the use of DataRobot. Past exposing me to a large range of strong strategies and algorithms to unravel complicated issues, it gives me with nice instruments to handle the entire life cycle of an AI answer, from knowledge preparation to MLOps.

Meet Belén Sánchez


Leave a Reply

Your email address will not be published. Required fields are marked *