Dhruv Bhatt | Head of Digital Transformation Office
Roche

Dhruv Bhatt is the Head of the Digital Transformation Office at Roche, Branchburg. With over five years dedicated to shaping the company's future, Dhruv has been instrumental in transforming traditional manufacturing and quality processes into a highly efficient, data-driven enterprise. His work centers on implementing a unified digital platform that leverages advanced technologies such as Snowflake, Tableau, and automated data pipelines, alongside applied and agentic AI that automate high-effort manufacturing and quality workflows. His focus is on harnessing data as a strategic asset to streamline operations, foster a culture of continuous improvement, and ultimately boost the company's efficiency and profitability.

Appearances:

Day 2: 29th October @ 14:20

From Hours to Minutes: Automating Batch Release Protocol with AI and Data Engineering

Batch release is one of the most quality-critical yet manually intensive steps in pharmaceutical manufacturing, historically requiring a multi-hour review across dozens of disconnected systems and documents. This session shows how Roche automated the Branchburg Batch Release Protocol using governed data pipelines and a fully deterministic document-generation app, compressing that review to under 30 minutes. Dhruv Bhatt also shares how AI-assisted development with Claude Code accelerated the build — tagging templates, writing data pipelines, and running audits — while the operator-facing runtime stays 100% deterministic, with every value traceable to its source. A practical blueprint for using AI responsibly in a validated environment.

Dhruv Bhatt, Head of Digital Transformation Office, Roche

Day 2: 29th October @ 14:40

Closing Panel: AI Reality Check: Navigating past the hype to deploy pragmatic, high-ROI AI architectures in modern R&D

What is the definitive operational hurdle a lab must clear to prove an autonomous agent is actually managing a LIMS workflow rather than simply automating a legacy bottleneck?
Should we prioritize universal data standards or invest in AI-driven translation layers that can retroactively capture the 'human context' missing from traditional ELNs?
How do we shift the burden of literacy?
Should we be training biologists to catch algorithmic hallucinations, or demanding that AI tools become natively 'scientist-literate' to protect critical thinking?

Yousef Baioumy, Automation Engineer II, Adaptive Biotechnologies