talk-data.com talk-data.com

Event

How We Use LLMs and Human Evaluations to Ensure High Quality of Suggestions

2024-09-26 – 2024-09-26 Meetup Visit website ↗

Activities tracked

1

Join us on Thursday, September 26, to explore our in-house solution for conducting human-powered quality evaluations at scale to gain insights into how new Grammarly features are evaluated before they are released to users.

✅ Registration: to attend the meetup, please register ➡️ here ⬅️

🔈 Speaker: Vitali Henne, Software Engineer

🚀 At Grammarly, the Engineering and Data teams have developed an in-house solution to conduct quality evaluations of suggestions as simply as possible. We use it to gain insights into the quality and impact of new features throughout our development cycle. In this presentation, we will review:

  • What are human quality evaluations, and why do we need them?
  • How we use this solution to get insights on the impact and quality of new features before they are deployed to production
  • A deep dive into the scalable, distributed design of the solution

Agenda: 18:30 Doors open: Time for mingling and networking with fellows; snacks and drinks will be served 19:00 Talk 20:00 More snacks, drinks, mingling, and networking 21:00 Meetup ends

✅ Where: In person, Grammarly Berlin hub ✅ When: Thursday, September 26 ✅ Language: English ✅ Use this link to register: https://gram.ly/3ZefJkk

The event is free. Registration is mandatory. Due to a limited number of seats, the invites will be sent to a limited number of registered on a first registered first invited basis. Please check your inbox for a confirmation email about your attendance.

Sessions & talks

Showing 1–1 of 1 · Newest first

Search within this event →

In-house solution for conducting human-powered quality evaluations at scale

2024-09-26
talk
Vitali Henne (Grammarly)

Overview of Grammarly's in-house solution for conducting quality evaluations of suggestions, including what human quality evaluations are, how the solution provides insights into the impact and quality of new features before deployment, and a deep dive into the scalable, distributed design of the solution.