What to do with a data scientist in your startup
Download as PPTX, PDF1 like92 views
A data scientist joins a startup and faces several challenges, including a lack of clearly defined goals, communication issues between teams, and not having access to data at the beginning. However, by getting involved in business logic and data modeling, the data scientist can help build the foundation for future data science work. It's important that startups clearly communicate expectations and allow data scientists to design a roadmap to data science work over time.
What to do with a data scientist in your startup
- 1. What to do with a data scientist in your startup The challenges and benefits
- 2. Hi there! I am Xanthippi. (Aspiring) Data scientist in OCL, working mainly with R. Graduated from Technical University of Crete, as a Production and Management Engineer with a specialisation in Operational Research. Agile and UX enthusiast. Main goal: To get to a point of combining math and coding for assisting in the development of a product, which is a result of iterations of UX research, and connecting it with the business needs and development. Side-goal : Gaining a better understanding of the world.
- 3. What about OCL? OCL it is a UK based startup. Building a “smart’s contract” framework, expressed as graph, which is distributable, decentralized and uses cryptography. “New institutions, and new ways to formalize the relationships that make up these institutions, are now made possible by the digital revolution. I call these new contracts “smart”, because they are far more functional than their inanimate paper-based ancestors. No use of artificial intelligence is implied. A smart contract is a set of promises, specified in digital form, including protocols within which the parties perform on these promises.” Nick Szabo, Smart Contracts: Building Blocks for Digital Markets
- 4. How clearly is DS defined? ● Data analyst ● (Big) Data Engineer ● ML engineer ● BI analyst ● Business analyst ● Researcher Source https://blog.datasciencedojo.com/data-science-skills/
- 5. A few challenges when joining a startup ● Not clearly defined goal/purpose ● Communication issues between business and tech team ● When is it going to be live? ● Funding concerns ● Not clearly defined job roles and not clearly defined needs of the startup ● Pivoting, aka throwing your work away
- 6. When a data scientist joins a startup.. Data science YOU THE TEAM ¯_(ツ)_/¯ ? source : https://imgflip.com/memegenerator/Skeptical- Baby
- 7. Challenges How to get to the data science part Where to start? How to fit in the team Challenges of a data scientist w/o data
- 8. Where to start You need DATA Business logic shall be integrated The team needs to “implement” all about DATA Becoming the DATA person source : https://www.zdnet.com/article/how-to-build-a-data-science- team/
- 9. Finding your way back to data science source : https://www.analyticsvidhya.com/blog/2016/12/21-reason-why-you-should-not-become-a-data-scientist/ So, far you: ● Are well aware of the business logic ● Have an overview of the data flow ● Were part of the data modeling You have been building the bridge to data science while walking it. So, take a step back, “zoom out” and start planning the reporting so you can get eventually to data science. Be patient, one step at the time. Treat it as a product, design a roadmap to get where you want.
- 10. Why you shouldn’t join a startup ✗ Dealing with uncertainty. ✗ Losing touch with data science. Possibly less than 30% of your time related to data science, as you consider it. ✗ What you are to deliver is a product, that requires the initial product to be built and get launched. ✗ You have to be flexible and adaptive. ✗ You have at first to become a generalist, while finding the time to become a specialist.
- 11. Why you should join a startup ✓ You have an overview of the project and product. ✓ You are well aware of the data modeling and data flow. ✓ You are well aware of limitations. ✓ You might be able to design with the team and implement an end to end data product as you envisioned it. ✓ You are becoming part of a bigger picture while expanding your understanding .
- 12. How a data scientist’s life could become easier Be well aware and communicate why you want a data scientist to join your team and what are your expectations. Allow them to design a roadmap, so they can get to the data science part the “right” way. The fun for them might start after the product has been launched. Communicate the business needs. Be aware that when the time comes they might need to focus on their thing and no longer be the data person. After all, the “data” thing it is not a one wo(man) show. :-)
- 13. At the end of the day..
- 14. Thank you! LinkedIn : www.linkedin.com/in/xanthippilemontzoglou Email : xanthippi.lemontzoglou@hotmail.com Twitter : @xlemontzoglou
Editor's Notes
- #3: Always aspiring there is a lot to learn
- #4: It has achieved that by building a related C framework for graph creation, the related tools and respective Web, node, swift and R APIs. These tools are being used for the development of specific software products and the framework is implemented in specific sectors, such as music industry and labels, that have a need for expressing and enabling licensing of artifacts for digital distribution and having knowledge of the related usage.
- #5: So once we finished with the introduction, let’s first talk a bit about the separate challenges of ds and startups. How many times they have approached you for a job that had the following job titles? So, even we as data scientists cannot be sure of what is expected of us..
- #6: A few challenges of a startup
- #7: Data science -No data -What to science? And the team tries to understand what you are doing?
- #8: You are a painter without colours and you have to fix your own
- #9: What you need to start is DATA What the see “in” you is a DATA person that is all about math, understands the business and knows a bit of coding. Data modeling -Performance / optimization -Integrating the Business needs in the data -Working with the people engineering the data -Knowing the data flows and learning the stack .. becoming the project’s “API” person Finding your place in the team
- #10: While doing so you are losing you touch, with data science or not? Taking a step back “zooming out” Check the goal and your target groups, start with reporting and then keep the answers and the problems to solve. It is a product with many stages of development, respect it, take time to plan
- #14: Overview with a dose of risk vs being an excellent gear of a broader “machine” . https://dyn.com/blog/culture-con-improving-intracompany-communication-by-uniting-specialists-generalists/