Daimler Truck is one of the world’s largest manufacturers of commercial vehicles, including trucks and buses. The company operates globally across four key regions—Europe, North America, Asia Pacific, and China.
Daimler Truck Innovation Centre India (DTICI) is a GCC based in Bengaluru, established three years ago.
It focuses on providing IT and engineering solutions for all segments, products, and regions of Daimler Truck. “Our goal is…to provide world-class engineering and IT solutions to our customers and products,” Raghavendra Vaidya, managing director and CEO of DTICI, told AIM.
The RAG Conversation
Vaidya also reaffirmed that retrieval-augmented generation (RAG) architecture remains a valuable approach. “I don’t think RAG architecture is dead. It’s working well for us,” the representative said. Companies can either retrain a model with their data or use RAG to enhance a GPT model. Both methods have their own advantages.
Even major companies like Microsoft support RAG by offering models that help vectorise data efficiently, as per Vaidya.
However, the conversation around RAG is not new.
RAG has revolutionised how AI systems process and respond to user queries by using external knowledge sources. However, as it doesn’t meet all the diverse needs of modern enterprises, everyone wants to replace RAG with something new.
This is where agentic RAG comes into play. Agentic RAG represents an advanced architecture that combines the foundational principles of RAG with the autonomy and flexibility of AI agents. It promises a future where AI systems are more adaptive, proactive, and intelligent.
Furthermore, last year, Google released its new Gemma model, DataGemma. While the world is experimenting with RAG to reduce hallucinations and increase accuracy, Google decided to use retrieval interleaved generation (RIG). This technique integrates LLMs with Data Commons, an open-source database of public data.
DTICI is not Building LLMs
Vaidya highlighted that DTICI is not building large language models (LLMs) but is currently using OpenAI’s LLM for internal purposes.
Machine learning (ML) remains a core focus, even though the term is less commonly used today, he further said. For over a decade, the company has been developing ML models from scratch, supported by a skilled team of data scientists and engineers who collaborate with business experts across different functions.
In the past year, the company elevated its approach by assigning accountability for AI and data initiatives to its Bengaluru team.
Regarding GenAI, DTICI is currently running multiple pilot projects to assess its potential impact. The company has already seen success in using Microsoft Copilot and GitHub Copilot to improve software development productivity, whether through code generation, test case creation, or code quality validation.
Beyond software engineering, DTICI is exploring GenAI in sales, procurement, and after-sales.
Rather than taking a technology-first approach, the company prioritises business needs. “We don’t bring in the technology, dabble and see what it can do. That’s not the approach we’re taking. We’re taking a business approach where we identify areas where it can produce business results and provide benefits, either to the top line or efficiency or the bottom line. And then we go and build a pilot around it,” Vaidya said.
DTICI also built a sandbox on Azure about a year ago, using an older version of OpenAI’s language model.
According to Vaidya, DTICI has been using the model for some time and finds it effective. It has created internal chatbots and assistants that use OpenAI’s language model and its own secure data.
Vaidya acknowledged that training a model with its own data would be better, but it would take excessive time and money. Instead, DTICI prefers its current approach and believes it to be a good balance.
DTICI’s Bengaluru Narrative
Vaidya said that Bengaluru remains the top choice for talent in India, with its unmatched depth and variety of skilled professionals. “The length, breadth, and depth of talent you have in Bengaluru is unmatched.”
He believes that as global capability centres (GCCs) grow, they may expand to other cities where talent is available. Some of them have successfully established operations in multiple locations. However, Bengaluru remains the first choice for new GCCs, and DTICI has no recent plans to expand to tier-2 cities.
At DTICI Bengaluru, the focus is on engineering and IT. The team develops intelligent software for trucks and buses. Most of the innovation and investment are happening in IT, software, and electronics. Highlighting this trend, Vaidya said, “If you want to increase the frequency of innovation, or you want to innovate faster, then I think software and electronics is the place to be.”
In IT, the company is deeply focused on using data, ML, and artificial intelligence.
Vaidya revealed that a major project of predictive maintenance, directed from Bengaluru, aims to predict part failures using analytics and ML instead of traditional physics-based methods.
The system analyses real-time truck data to forecast when a part is likely to fail. However, accuracy is critical for this to be effective. “If the model is not 85% or more accurate, then nobody is going to buy it,” Vaidya said.
Since customers rely on these predictions to replace parts before failure, achieving high accuracy is essential. DTICI has been deploying these solutions over the past few years, and they have proven extremely effective in terms of profitability and cutting warranty costs.
“It is pretty simple; you get to work on the bleeding edge of the technology and the work that you do makes either a product better or customers more profitable,” Vaidya concluded.
