The Test Tribe Mega Meetup - Bangalore
Bangalore, we’re meeting for the 58th time, and this one’s a bigger gathering than our usual monthly meetups.
If you’ve been to one of our meetups before, it’ll be good to see you again. If this is your first, you’re very welcome. Come for the talks, and take the chance to meet more people from Bangalore’s testing community.
See you on 28 November.
Details
Agenda
About the event
Session 1
Topic - Token Optimization Techniques For AI Assisted Development
AI-assisted coding is now part of everyday engineering work, and token costs are quietly becoming a real budget problem. Some teams have already cancelled AI tool licences, and others have used up a full year’s AI budget in a few months. This session shows practical ways to cut token spend while keeping AI responses just as good. Using hands-on experiments with GitHub Copilot and other AI agents, we’ll cover how to keep context lean, write shorter prompts that stay precise (with before and after examples), limit unnecessary output, send each task to the right AI mode, pick the right model for the job, and avoid mid-session changes that break caching and drive up costs.
We’ll also walk through an open source Token Efficiency Toolkit, where each tool tackles a different source of token waste. The demos use GitHub Copilot in VS Code, but the techniques work with most AI coding agents.
Key Takeaways
Everyday techniques to cut AI token use
Spend fewer tokens, get the same results
Speaker: Dimpy Adhikary, Staff Quality Architect at Vonage
Dimpy has over 20 years of experience in Quality Engineering, leading teams across test architecture, automation and performance engineering. At Vonage, she is leading a Voice Modernization project, moving legacy test frameworks to a modern automation stack for mission-critical systems. She explores how AI agents can make engineering more efficient in practical, responsible and cost-effective ways. Dimpy is a past conference speaker and an active mentor in the QA community.
Session 2
Topic: Observability in AI Testing: From Execution Tracing to Quality Evaluation
Observability gives testers clear visibility into how AI systems behave in real use. End-to-end tracing follows each request from the user prompt through retrieval, model calls and tool usage to the final response, so it's easy to see exactly where something breaks. Performance monitoring tracks latency, token usage, cost and error rates, helping teams catch slowdowns and regressions early.
It also helps teams judge the quality of AI output. Failure analysis uses traces to find patterns behind hallucinations, off-topic answers and broken tool calls. Response evaluation, through automated scoring, LLM-as-a-judge or human review, turns real user interactions into ongoing test data, giving QA teams steady confidence in their AI systems.
Key Takeaways
See every step of an AI request
Test quality as well as uptime
Speaker: Sandeep Dora, Lead Software QA Engineer at Schneider Electric
Sandeep Dora is a QA leader with 12 years of experience testing AI and ML applications, including Generative AI, chatbots, agentic AI and RAG systems, along with web, mobile, desktop and OTT platforms. He specializes in AI model validation, prompt and response testing, and end-to-end data pipeline QA, with hands-on work in API testing, Selenium and Python automation, and AWS cloud testing. Sandeep has led QA teams in Agile environments across media and entertainment, retail, e-commerce and life sciences, with a focus on making AI systems accurate, reliable and ready to scale.
Who should attend
Testers, QA engineers, SDETs, automation engineers, quality leaders, and anyone building or breaking software for a living.
A look back
We've done this before, just smaller. Here's what our monthly meetups usually look like - expect more of this, at a bigger scale.
About The Test Tribe:
The Test Tribe, established in 2018, is the world's largest software testing community and an EdTech startup. We empower testing professionals globally with expert courses, memberships, events, and community initiatives, fostering collaboration, learning, and growth. With over 980+ events and a global reach of 175K+ testers from 130+ countries, our mission is to provide life-altering growth to every testing professional.