Panel: ML ecosystem in Brazil
summary
Brazilian ML scene faces concrete frictions: hiring senior talent is scarce while juniors are abundant; production is the bottleneck — building models is easy, first mile data / last mile delivery and serving models reliably is hard; large firms report legacy systems, legal/privacy limits and a high failure rate where 20 ideas → 1 product; startups cannot afford failures because models change reality so deployment must be precise; universities struggle with one year bureaucracy delay to partner with industry, though funded postdocs and one-day-a-week publication policies are helping.
AI-written summary · about
chapters (5)
- 0:00Introduction and panelists
- 2:58Biggest ML challenges today
- 14:45University-industry collaboration
- 22:00Breaking into data science
- 28:51Closing remarks and Q&A
related talks
46mPanel: OpML
42mDeploying ML models and all the things that go wrong
21mWhy is Production ML so Hard?
30mThe European AI Ecosystem
32mTeam Teardown: Evolving an ML Platform
59mWhy Great Models Fail: Lessons From 9 Years of Deploying ML Models
44mAI workflows and teaming - how do teams deliver AI products?
37mPanel: Data Analytics In Investment Banks