On this page, we’ve said many times that the car is no longer just a machine built from mechanical parts. Sensors, software, data and algorithms are becoming as fundamental to a vehicle as brakes, steering and the powertrain. But what does this mean for automotive engineers? Caleb Munigety, head of enterprise AI at Bosch India, told us on our podcast last week that it means a career built around one discipline will increasingly have to stretch across several.No, young engineers don’t have to become experts in everything. But they must go deep in one discipline and have a broad enough understanding of how the other areas connect. Caleb, who holds a doctoral degree from IIT Bombay, said this T-shaped skill set will become increasingly important as AI changes automotive engineering.The change is already visible in something as basic as braking. Traditionally, braking was largely a mechanical and electronic problem. Today, sensors collect data, software analyses it, and algorithms can determine how the system should respond. In an electric vehicle, braking can also generate energy that is returned to the battery – thanks again to algorithms. Similar changes are transforming steering, powertrains and safety systems.“The idea is that if I am able to sit with the right engineering team and start asking the right questions, that’s sort of an integrated view,” Caleb said.An experienced automotive engineer, Caleb said, does not suddenly need to become a data scientist or neural-network developer. Instead, the engineer should understand enough about AI to work effectively with specialists and challenge what AI produces.He gives the example of a vehicle-dynamics engineer with years of experience. If an AI system produces an algorithm to optimise vehicle behaviour, the engineer does not need to write the neural network. But the engineer needs to know what questions to ask, what the system can and cannot do, and whether its answer makes engineering sense.
The journey of a future mobility engineer, or for that matter any engineer, has to start with knowing the fundamentals in that domain. The second step is, because the world is increasingly digital and AI, they should build a certain fluency in AI, data systems, software systems. The third important thing is, have real-world experiences. That means they need to talk to people in the field, go for internships … People often gather certificates. But learning is one thing, doing is another. They must build, test, fail, correct, improve. If one can go through this journey, then their employability will be higher when people like us hire.
Caleb Munigety, Head, Enterprise AI, Bosch India
In fact, Caleb argued that traditional engineering fundamentals will become more important, not less. An AI system might generate a thermal-management design, for instance, but a domain expert has to ask whether it complies with thermodynamic principles, whether it will work after years of ageing and how it will perform in extreme conditions.“The judgment call finally should lie with the automotive engineers,” he said. “It has produced something, but is this really engineering?”That puts a premium on what he called AI and data literacy. At Bosch, he said, the aim is not to turn every domain expert into an AI engineer. Instead, engineers are given curated content that helps them understand what AI is producing and make decisions about it.AI’s touching entire engg lifecycleCaleb said AI is touching the entire engineering lifecycle. AI can analyse large volumes of requirements from engineering teams, customers and regulators. It can then generate design alternatives, optimise configurations and improve simulations.During manufacturing, AI can be used for quality inspection and predictive maintenance. For autonomous vehicles, it can generate synthetic data for rare “corner cases” that would be difficult or dangerous to collect in the real world. AI can also generate software code and, after vehicles are on the road, use data from connected vehicles to improve their performance.The biggest transformation, he said, will come when all these capabilities connect rather than operate as isolated tools.Caleb encourages students to use simulation before rushing to physical prototypes. Simulation allows them to test a large number of alternatives cheaply, narrow the options and then build physical prototypes around the most promising designs.Electric vehicles offer another obvious area for young engineers. Caleb said students should understand battery systems and, more importantly, how AI can make battery management intelligent.But his advice is not to abandon internal-combustion technology overnight. Hybrids and the gradual transition to electric mobility mean engineers still need to understand the fundamentals across technologies.
Source: timesofindia.indiatimes.com
