Highlights

Budapest Data + AI Forum 2026: How to Engineer Reliability at Scale

I spoke at the Budapest Data + AI Forum 2026 on why ML deployments fail in production and how to engineer reliability at scale. Session abstract, key takeaways, and links.

I had the privilege of speaking at the Budapest Data + AI Forum 2026, one of Central Europe's largest data and AI conferences, held May 18-20, 2026 at the Danubius Hotel Helia in Budapest as a hybrid event.

My session, "Why ML Deployments Fail in Production and How to Engineer Reliability at Scale," looked at a problem most organizations underestimate: the gap between a model that works beautifully in research and a model that stays reliable once it meets real infrastructure, data pipelines, and production traffic.

About the session

Machine learning breakthroughs get the attention, but the harder, less glamorous challenge is deploying those models reliably into production. Models that perform well in research environments frequently fail once they interact with real-world infrastructure, data dependencies, and production workloads.

Production ML systems have to coordinate across many layers: training pipelines, data dependencies, deployment infrastructure, and serving systems. Small gaps in readiness across any one of them lead to failed deployments, unstable rollouts, or degraded performance. Drawing on real platform-engineering experience, the talk explained why production ML is often harder to operate than traditional software, and what teams can actually do about it.

Key takeaways

Why machine learning deployments frequently fail in production environments
The hidden system dependencies that complicate ML deployment pipelines
How reliability challenges differ between traditional software systems and ML systems
Platform-engineering strategies that improve deployment success rates
Practical approaches for enabling safer, more reliable model rollouts

Links

Speaker profile and session: Ankur Gupta, Budapest Data + AI Forum
Event website: budapestdata.hu
Detailed schedule: Forum schedule
AI & ML speaker lineup: Speakers

If you attended the session, or would like the slides or a recording, feel free to reach out. Always happy to talk about production ML reliability.

Characterizing GPU Capacity and Computational Cost in Production AI Inference, IEEE MetroInd 2026, Rome

IEEE MetroInd 2026: Characterizing GPU Capacity and Computational Cost in Production AI Inference

My paper on a replica-centric capacity-planning framework for GPU and Multi-Instance GPU (MIG) based AI inference platforms was accepted at IEEE MetroInd 4.0 & IoT 2026 in Rome, Italy.

How automation is changing managerial roles

I recently sat down with Business Standard for their Manager's Mantra series (BSmart). What started as a conversation about automation and managerial roles turned into a wide-ranging discussion: enterprise cloud strategy, the economics of AI infrastructure, how to build data platforms people actually use, modernizing legacy systems,

IEEE CSNT 2026: Operationalizing Site Reliability in Large-Scale Distributed Systems

I presented our paper "Operationalizing Site Reliability in Large-Scale Distributed Systems: Shifting Ownership Left" at IEEE CSNT 2026 (Al-Khobar, Saudi Arabia): predictive models that forecast failures to keep large-scale distributed systems reliable.

Ankur Gupta on Zee Business India 360, AI Infrastructure Expert, 18 Feb 2026

On Zee Business: Why AI Hallucinates, and How to Use It Responsibly

I joined a Zee Business India 360° panel (Feb 2026) on AI hallucinations: why models confidently make things up, how to verify their output, practical prompt techniques, and where AI regulation stands across the EU, US, and India.