Sonika
Janagill
Lead Engineer at VML
"Building Production AI with Gemini: Beyond Tutorials to Real-World Systems"
"Building Production AI with Gemini: Beyond Tutorials to Real-World Systems"
Wed May 13 - 4:10 AM EDT/New York (See in local time)
Add to Calendar
05/13/2026 4:10 AM
05/13/2026 04:50 AM
America/New_York
#WTGC2026
"Building Production AI with Gemini: Beyond Tutorials to Real-World Systems"
#WTGC2026
"Building Production AI with Gemini: Beyond Tutorials to Real-World Systems"
https://www.womentech.net/ringcentral
https://www.womentech.net/ringcentral
"Building Production AI with Gemini: Beyond Tutorials to Real-World Systems"
"Building Production AI with Gemini: Beyond Tutorials to Real-World Systems"
Don’t miss out and join visionaries, innovators, and thought leaders from all over the world at the Women in Tech Global Conference.
Vote by Sharing
Unite 100 000 Women in Tech to Drive Change with Purpose and Impact.
Do you want to see this session? Help increase the sharing count and the session visibility. Sessions with +10 votes will be available to career ticket holders.
Please note that it might take some time until your share & vote is reflected.
Session: Building Production AI with Gemini: Beyond Tutorials to Real-World Systems
Getting your first Gemini API call working feels like magic. Shipping it to real users feels like defusing a bomb you built yourself.According to Gartner, 95% of generative AI projects fail to deliver measurable ROI. Not because the technology does not work, but because there is an enormous gap between "the demo works on my laptop" and "this runs reliably in production." Most tutorials take you to the first milestone and wave you off cheerfully. This session covers the rest of the journey.
I will walk through a real multi-agent system built with Google's Agent Development Kit, showing how different Gemini models, Flash, Pro, and the Live API, play different roles within the same architecture, like a well-run kitchen where not every chef needs to be Gordon Ramsay. Some tasks need speed, some need reasoning, and choosing the wrong one costs you time and money.
The second half is the bit nobody puts in their quickstart guide: what happens when one agent hands off to another and something quietly goes wrong? How do you actually know what your production agents are doing? And how do you test an AI system before your users become your QA team?
You will leave with a working mental model, a production checklist, and a healthy sense of what "production-ready" actually means, beyond the tutorial page.
Key Takeaways
- Multi-agent systems without the mystery: A clear, practical look at structuring real systems where multiple Gemini-powered agents collaborate, using Google's Agent Development Kit as the foundation
- Gemini model selection that saves you money: Concrete criteria for choosing Flash versus Pro versus the Live API, because using the wrong model for the job is like hiring a surgeon to change a lightbulb
- Stopping failures before they cascade: The error handling patterns that prevent one agent's bad day from becoming an expensive, hard-to-debug disaster across your whole system
- Observability for systems that think: What to log, what to trace, and what to alert on, so you are not flying blind once your agents are live
- A production checklist you can use immediately: Evaluation strategies, cost controls, and the practical steps that separate a working demo from something you can ship with confidence
Bio
Sonika Janagill is a Lead Backend Engineer and Engineering Advocate whose 19-year career tells a story of continuous transformation. From building enterprise Java systems for 14 years to winning AI hackathons, implementing production agentic systems on Google Cloud, and earning both the GCP Professional Architect and Google ML Engineer certifications, she's living proof that it's never too late to reinvent yourself in tech.What sets Sonika apart is her commitment to bringing others along on the journey. Through the Women Coding Community AI Learning Series and the VML AI Guild, she mentors aspiring developers and runs global knowledge-sharing sessions on AI agents and production cloud systems. She is a Google Developer Expert (GDE) in Cloud AI.
She's passionate about making complex AI technologies accessible to engineers everywhere, regardless of where they are in their journey. The best time to learn something new is always now.
