Welcome to the JS Meetup #101. Yes, we crossed 100 sessions in this community.
I invite you to the meetup which keeps pace with emerging technologies. We will have a mix of topics this meetup. More talks are getting finalized.
Agenda:
10:00AM Welcome & Introduction to meetup
10:10AM Talk 1: Mastering LLMs: Fine-Tuning, Prompt Engineering & Steerability; Speaker: Shreyas chandavarkar
10:50AM Talk 2: Node.js’s Native Testing: A Beginner-Friendly Introduction; Speaker: Nabil Nalakath
11:30AM BREAK
11:50AM Talk 3: Build applications with GitHub Copilot agent mode; Speaker: Tadit Dash
12:40PM Photos and Networking
01:00PM Close of event
Talk 1: Mastering LLMs: Fine-Tuning, Prompt Engineering & Steerability
Speaker: Shreyas chandavarkar
Abstract: Large Language Models (LLMs) are powerful, but getting them to consistently follow complex instructions remains a challenge. Fine-tuning offers precision but is costly and rigid, while prompt engineering is flexible but unreliable. Neither approach fully solves the problem of steerability—an LLM’s ability to understand, adhere to, and interpret complex system prompts. This talk explores the trade-offs between fine-tuning and prompt engineering, helping attendees develop the thought process required for Generative AI Engineering. We will discuss techniques like system prompts, retrieval-augmented generation (RAG), and hybrid fine-tuning to improve instruction adherence. Additionally, we will cover strategies to reduce costs while maintaining performance, including quantization and model distillation. Attendees will gain practical insights into making LLMs more reliable, controllable, and cost-efficient, ensuring better AI-driven applications.
Key Takeaways: Fine-Tuning vs. Prompt Engineering: When to use which Steerability as a key metric for LLM development Techniques to improve LLM adherence to complex prompts Cost-efficient methods for optimizing AI models
Talk 2: Node.js’s Native Testing: A Beginner-Friendly Introduction
Speaker: Nabil Nalakath
Abstract: In this beginner-friendly talk, we’ll walk through how to set up and effectively use Node’s native test runner—without relying on external libraries like Jest or Mocha. We’ll explore test structuring, asynchronous patterns, and best practices for maximizing the benefits of this lesser-known feature. By the end of the session, you’ll have a clear roadmap for integrating Node’s built-in test runner into your projects and simplifying your testing workflow.
Talk 3: Build applications with GitHub Copilot agent mode
Speaker: Tadit Dash
Abstract: GitHub has introduced a groundbreaking feature to its Copilot offering, previewing the agent mode. This innovative feature aims to revolutionize code editing workflows by allowing developers to leverage natural language prompts and automate repetitive tasks. With agent mode, Copilot can autonomously plan and execute tasks, suggesting and running terminal commands, and iteratively resolving issues.
---
We have listened to your feedback. We are incorporating below:
a) We are putting a cap on RSVP, so no separate online form needs to be filled; which was confusing in previous meetups. If meetup.com shows RSVP=YES, then you are eligible to attend.
b) We will maintain start time and be punctual in future events.
