Python talk night at City University Seattle (Belltown)
Details
Join us for an exciting talk night at City University of Seattle's Belltown Campus!
Agenda:
5:30-6pm: Doors open/networking
6-6:10pm: Opening remarks
6:10-6:35pm: Talk #1
6:35-6:50pm: Intermission
6:50-7:05pm: Lightning talk
7:05-7:30pm: Talk #2
7:30pm-8pm: Networking
After party to follow at Teku Tavern (552 Denny Wy, Seattle, WA 98109)
What we'll do:
This event will feature the following presentations:
Talk #1: LLM-powered Merge Conflict Resolver
Speaker: Advitya Gemawat
Description: Merge Conflicts in software programming occur in 20% of all merges in open-source projects, incurring a median resolution time of ~6 minutes for a single conflict at best, or multiple days (from conflict detection to incorrect resolution to fixing downstream feature regressions) at worst, especially in large-scale software projects. Drawing inspiration from the scale and complexity of merge conflicts in one of the planet's largest codebases (the Windows OS repo at Microsoft), this talk will touch upon a variety of insights across crafting custom LLM Evals, Azure OpenAI Fine-Tuning, Qualitative Insights, and most importantly, the usage of multiple Python SDKs and best practices to craft a scalable LLM-powered Merge Conflict Resolver adapted towards codebase-specific conflict resolution patterns. Similar to how LLM Coding Agents transformed the SWE workflow from writing code to generating & reviewing code, this project is a natural evolution of the manually-driven Merge Conflict Resolution process in terms of reviewing LLM-generated conflict resolutions. The Fine-Tuning and Evals aspect of adapting the LLM towards a particular codebase makes the approach unique and better performant (>2x better resolution accuracy) than the AI-based conflict resolution capabilities with vanilla LLMs offered in most IDEs.
Lightning talk: TBD
Speaker: William Harding
Description: TBD
Talk #2: TBD
Speaker: TBD
Description: TBD
What to bring:
Since this is a talk format meetup, you don't need a computer.
All attendees are required to abide by the PuPPy Code of Conduct, you can read the Code of Conduct here: https://www.pspython.com/code-of-conduct/
Building access:
Please enter the building at the 6th Ave and Wall street entrance.
Parking:
The only parking for the event is street and paid garage parking.
Transit:
City U's campus is within walking distance of numerous buses including stops on 5th and Wall St, 3rd Ave and Bell St, 3rd Ave and Cedar St, 7th Ave and Denny Way and Dexter Ave and Denny Way.


