Sep 09: Deloitte announced the launch of its Open Model Engineering practice in markets around the globe. The practice will initially be launched in India and other key global markets. The move expands Deloitte’s AI services portfolio in response to an evolving enterprise landscape, where competitive advantage is shifting from model intelligence to accelerating enterprise deployment, re-designing organisational workflows and the coupling of human judgement with agentic efficiency.
As more and more enterprises and governments increasingly explore open AI models as a way to gain deployment flexibility, manage token costs, and increase choices through a mixture of models, the new practice aims to help organisations design, engineer, deploy and operate, and scale AI solutions using open models and full stack open-source technologies tailored to each client’s business requirements, regulatory environment and technology footprint.
“Enterprise AI is becoming an architecture choice, not a model choice. Organisations will need the freedom to move workloads across models and environments as economics, regulation and business needs change. Open models add an important dimension to that strategy because they allow enterprises to shape AI around their own data, processes and context rather than designing the enterprise around the limitations of a model. Our Open Model Engineering practice is designed to build that flexibility in from the start, without compromising governance or the distinctiveness of the business,” said Sathish Gopalaiah, President, Consulting, Deloitte South Asia.
According to Deloitte, enterprises today are navigating four important considerations in AI deployment: flexibility to select the right model and architecture for each workload, predictable cost economics as Agentic AI usage scales, sovereignty over where and how models are deployed, and control over enterprise data, intellectual property (IP), and model behavior with transparency on inference.
Through the Open Model Engineering practice, Deloitte aims to help clients make model and architecture choices across proprietary models, open models, Agentic platforms, cloud services and/or on-premises infrastructure; manage token economics by choosing the right model, infrastructure and deployment patterns to improve return on investment; and provide greater flexibility to build AI applications of open models and further fine-tune them for geographical, linguistic and cultural context. The practice also helps clients gain greater choice over where models run, what data is uploaded, and how competitive advantage is retained by protecting IP and the uniqueness of generated inference.
Agentic AI capabilities powered by NVIDIA Nemotron family of open models Deloitte’s Open Model Engineering practice will initially focus on developing enterprise AI applications built on NVIDIA Nemotron open models and NIM microservices. Deloitte firms will also help clients build Sovereign AI stacks using open-source AI frameworks, fine tune open models, and enhance cyber defenses with open harnesses.
“Enterprises need both open and proprietary models for different workloads,” said Kari Briski, vice president of generative AI at NVIDIA. “Deloitte’s Open Model Engineering practice, built on NVIDIA Nemotron models, gives developers greater control over enterprise data and proprietary information while enabling secure, responsible deployment of Agentic AI.”
As part of this practice, Deloitte firms will leverage the Zora AITM digital workforce platform, with agentic harness capabilities built on NVIDIA Nemotron open models and NIM. This gives clients a path to deploy, continuously optimise and govern AI agents using open models within their own environments.
Investing in talent: Forward deployed engineers Deloitte plans to hire, train, and certify forward deployed engineers globally certified in open models. These specialists will work directly alongside Deloitte firm clients to implement open model solutions and will serve as the operational foundation of the new practice.