Sanghamitra Goswami at LeadDev West Coast 2023
summary
Claims that a small team moved from “scrappy to scale” in 3–6 months, leveraging existing investments: a data Lake and an ML pipeline that cut model-to-prod time. Numbers: five ML features in prod by end of 2022 and one ML noise‑reduction model since 2019. Built three generative features—status updates, postmortem drafts, and process‑automation code generation—using multiple LLMs, early access pilots, hack weeks, legal/security gating, and a human in the loop for safety.
AI-written summary · about
chapters (8)
- 0:00PagerDuty introduction
- 1:46AI landscape before generative AI
- 5:18Generative AI challenges
- 7:26Multimodal generative AI adoption
- 13:07LLM selection and principles
- 15:38Key takeaways and trust
- 16:43Generative AI features
- 21:07Closing and Q&A
related talks
38mFrom scratch to production in three months: the AI-first delivery model
28mPractical steps to establish and scale your data & ML team
10mFrom Model to Production: Building Scalable MLOps Pipelines
32mTeam Teardown: Evolving an ML Platform
21mImproving your Developer Experience Through Platforms and Generative AI
22mAccelerating ML impact
29mSeveral People Are Tuning: Data and Machine Learning at
18mBuild like it's 2030. Listen like you mean it.