Going Full Stack with Data Science: Using Technical Readiness...
summary
Argues for using NASA’s Technical Readiness Level (TRL) scale to move data work from notebooks into production, with concrete checkpoints: TRL-like milestones, targeted stakeholder engagement, and who must be involved at each stage. Gives a hate-speech ML example moving from Jupyter notebook proof‑of‑concept to an MVP integrated with front end, QA, legal and ops. Describes a real tech stack (Postgres, Redshift, Kafka/Scala/Python), emphasizes iterative failures, on‑call implications, and hiring to close specific readiness gaps.
AI-written summary · about
chapters (10)
- 0:00Full-stack data science intro
- 2:55Hate speech platform example
- 4:00Defining full-stack in DS
- 8:13Data science in tech stack
- 10:23DS in critical system path
- 14:00Full-stack DS team model
- 16:31TRLs for data science
- 20:02TRLs for hate speech example
- 24:40Key takeaways for DS teams
- 29:20Summary
related talks
28mPractical steps to establish and scale your data & ML team
20mWhy You Should Think Like an End-to-end Data Scientist, and How
17mProductionizing Data Science Workloads
26mCompleting the machine learning loop
28mDataEngBytes 2023 - PER-T1-05 - Ellie Gholami
25mBuilding A Successful Data Science Team
27mA Look Inside Applied Machine Learning
28mLeading machine learning research journeys with happy (into-production) endings