In this episode, Tanisha Bansal explains what it takes to build reliable AI agents for enterprise: from LLM evaluation, cloud infrastructure, authentication, and monitoring to human oversight and cost control.

She also shares why a successful demo isn’t production-ready, how AI is changing hackathons and technical interviews, and how community, mentorship, and speaking helped shape her engineering career.

Connect with us:

Tanisha Bansal - https://www.instagram.com/grow.with.tanisha

Animesh Pathak - https://linkedin.com/in/sonichigo

Women in tech are still too often told they have to choose between engineering, community, and leadership - Haimantika proves you can build all three into one powerful career. She also opens up about the “silly” side projects that quietly became career accelerators - including a GitHub Roast app that went viral, a Gen Z translator, and an AI plank coach built to solve her own workout problem. Those experiments weren’t just for fun. They became the fastest way to learn new tools, understand how products really work, and create proof that she could go far beyond tutorials.

You'll discover why DevRel is not a step away from engineering, but often the place where engineering, product sense, content, and community all collide. Haimantika breaks down the three C’s of DevRel - core, content, and community - and explains why the future of the role is even more technical in the age of AI, not less. We also get into the reality of AI-era building: why shipping is easier than ever, but understanding systems, architecture, benchmarking, and production tradeoffs matters more than ever. She shares what companies and communities can do better for women in tech, from making events feel safer and more welcoming to actively bringing more women speakers, engineers, and builders into the room.

Haimantika is a developer advocate at DigitalOcean, formerly in DevRel at Appwrite and a software engineer at Microsoft. She is also the author of an ebook on DevRel for startups and brings a rare mix of technical depth, community perspective, and honest career reflection. Perfect for developers, aspiring DevRel professionals, women in tech, and anyone who wants a sharper view of where engineering is headed next.

In this episode of Rollback Café, Sylvian joins us to explore how AI is reshaping incident response and modern SRE practices.

From faster detection to smarter debugging, we unpack where AI is actually adding value-and where it still falls short in real-world production environments.

We discuss:

• how AI is being used in incident detection and response
• the role of prompting in debugging and investigation workflows
• opportunities and limitations of AI in high-pressure incidents
• how SRE teams can integrate AI without compromising reliability
• the evolving skillset required for engineers in AI-assisted operations

This conversation focuses on practical, experience-driven insights-not hype-around applying AI in reliability engineering.

Who this is for
SREs, DevOps engineers, platform teams, and engineering leaders working on incident management, reliability, and production systems.

Connect with us on LinkedIn:

• Sylvain Kalache: https://www.linkedin.com/in/sylvainkalache/
• Animesh Pathak: https://www.linkedin.com/in/sonichigo/

Subscribe for more conversations on DevOps, platform engineering, and real-world software delivery.

#SRE #IncidentResponse #AIEngineering #DevOps #PlatformEngineering #ReliabilityEngineering #TechPodcast #RollbackCafe