Advancing your data science career
•
0 likes•212 views
The document provides career advice for advancing in a data science career. It recommends focusing on infrastructure skills like AWS and Kubernetes ("get into infrastructure"). It also advises being product-oriented by helping product managers, talking to stakeholders, and taking end-to-end ownership ("be product-oriented"). Additionally, it suggests using the 80/20 rule to focus on the most important 20% of skills that provide 80% of results.
Report
Share
Advancing your data science career
- 1. Advancing Your Data Science Career Alexey Grigorev Lead Data Scientist at OLX Group Founder at DataTalks.Club
- 3. �� Me when I started in DS How to make sure my projects are useful? Should I work on something because it’s cool? How do I get my models in production?
- 4. Just knowing machine learning is not enough*
- 5. 🔧 Me working at a startup How to prioritize because there’s so much to do and so little time? How to build this data/ML platform?
- 6. Lead Data Scientist at OLX Group Making sure ML projects are getting adopted by ● Mentoring others ● Scoping projects Me now
- 7. Career tips ● Get into infrastructure ● Be product-oriented ● Use the 80/20 rule
- 10. Mostly engineering work! Hidden Technical Debt in Machine Learning Systems https://papers.nips.cc/paper/2015/file/86df7dcfd896fcaf2674f757a2463eba-Paper.pdf
- 11. Career tips ● Get into infrastructure ● Be product-oriented ● Use the 80/20 rule
- 12. Get into infrastructure ● Learn AWS / GCP / Azure / whatever ● Deploy models yourself ● Ask for help ● Learn Kubernetes / Terraform / whatever
- 13. Career tips ● Get into infrastructure ● Be product-oriented ● Use the 80/20 rule
- 14. Picture: CRISP-DM The most important step
- 15. Be product-oriented ● Help your PM ● Talk to stakeholders ● Ask “why?” ● Take end-to-end ownership
- 16. Career tips ● Get into infrastructure ● Be product-oriented ● Use the 80/20 rule “Full-stack data scientist”
- 17. Becoming full-stack Skills ● Product management ● Data analysis ● Data engineering ● Backend engineering ● DevOps ● Setting up wifi in the office 🤯
- 18. Accuracy Data
- 25. Skill proficiency ● OK — can use the skill, mostly independently ● Good — can use the skill independently ● Great — expert
- 26. We don’t need to be experts in everything!
- 35. Areas Depth Product Management Data Engineering Machine Learning Backend Engineering DevOps Me now
- 36. Areas Depth Product Management Data Engineering Machine Learning Backend Engineering DevOps How other DS see me * * maybe not
- 37. Areas Depth Product Management Data Engineering Machine Learning Backend Engineering DevOps How engineers see me * * maybe not
- 40. Career tips ● Get into infrastructure ● Be product-oriented ● Use the 80/20 rule “Full-stack data scientist”
- 42. 80/20 rule ● Break down the role into core areas and skills ● Order the skills by importance ● Pick the most important ones ● Practice practice practice
- 43. MoSCoW method ● Create a list of features/ideas ● Mark each as “must-have”, “should-have”, “could-have” ● Take one (or two) the most impactful must-have ones ● Make it work with least effort ● Iterate
- 44. DevOps ● Infrastructure ● Automation ● Monitoring ● Reliability
- 45. DevOps ● Infrastructure ● Automation ● Monitoring ● Reliability ● AWS ● Kubernetes ● Terraform
- 46. Help! ● What’s most important? ● What to select? Find a mentor! (<== Bonus tip!!!)
- 48. Career tips ● Get into infrastructure ● Be product-oriented ● Use the 80/20 rule ● Find a mentor and join a community “Full-stack data scientist”
- 50. Backup
- 51. Product management ● Strategy ● UX & Design ● Communication ● Planning ● Evaluation
- 52. Product management ● Strategy ● UX & Design ● Communication ● Planning ● Evaluation ● Requirement gathering ● Prioritization ● Stakeholder management
- 53. Data engineering ● SQL Databases ● NoSQL Databases ● Stream processing ● Batch processing ● Data pipelines
- 54. Data engineering ● SQL Databases ● NoSQL Databases ● Stream processing ● Batch processing ● Data pipelines ● MySQL ● AWS (S3, Kinesis) ● Spark ● Airflow
- 55. Backend engineering ● SQL & NoSQL Databases ● CS fundamentals ● Languages and frameworks ● Web services ● Best practices
- 56. Backend engineering ● SQL & NoSQL Databases ● CS fundamentals ● Languages and frameworks ● Web services ● Best practices ● Python ● Docker ● Tests ● Clean code