Siddhish Sutaria and Jolly Shah: Shaping Embedded System Evolution
Siddhish Sutaria, Staff Software Engineer at Apple, and Jolly Shah, Staff Software Engineer at Google, are at the nexus of intelligent embedded engineering. Sutaria’s work drives innovations in audio processing and embedded ML for widely used devices such as Apple AirPods Pro and Vision Pro, blending advanced consumer-facing features with robust firmware design.
Shah, meanwhile, leads power management and firmware development for large-scale data center infrastructure and sensor-rich platforms, ensuring that intelligence at the device level can scale securely and efficiently across billions of endpoints worldwide. With combined experience spanning over three decades, their collaborative insights offer a rare look into how dual-domain expertise in consumer electronics and cloud-scale firmware design is setting new standards for intelligent product architecture and end-user impact.
Their professional achievements are deeply grounded in the steady progression of embedded technology. Both engineers understand the responsibility inherent in creating devices that touch billions of lives, balancing the pursuit of customer-centric features with the operational rigor required for scalable, sustainable system performance. As global regulatory shifts and accelerated AI adoption reshape the experiences consumers expect from devices, Sutaria and Shah’s partnership provides a template for embedded systems leadership that cuts across company lines and product ecosystems.
Vision and responsibility
Advancements in embedded intelligence are guided by a shared sense of purpose. “Two visions drive the work. One is customer first. Lots of intelligent AI systems end up in the hands of customers. What is important is how that improves customers’ lives,” Sutaria explains. This commitment is matched by a sense of scale: “Working on systems which get shipped in millions of units and used by billions of customers comes with lots of responsibility. We have to make sure products are engineered properly and executed in time.”
The regulatory environment increasingly demands rigor in engineering as well as customer focus. The recent requirements outlined by the European Commission’s Digital Markets Act mandate that platforms like Apple’s must ensure third-party access to core OS features without friction.
“We always want a more powerful system at the end of a product cycle,” Shah notes, underlining the dual imperative to innovate and to scale reliably as customer expectations and regulatory oversight intensify.
Complementary domains
Embedded engineering now exists at the intersection of device intelligence and system-level efficiency. “Lots of audio intelligence and AI needs large scale infrastructure to run it. Most AI runs on the cloud, and some of the AI runs locally on the device.” Sutaria observes.
This bidirectional influence means device teams must ensure hardware can leverage remote compute power without draining energy, while infrastructure architects like Shah ensure their systems empower the proliferation of increasingly intelligent, connected consumer products. The ecosystem convergence is facilitated by advances in connectivity and interoperability.
The mandated rollout of cross-platform data portability tools between Apple and Google underscores how infrastructure decisions influence user experience, while also supporting new device-to-cloud workflows. “There are domain-specific elements and some issues that overlap, but it’s all embedded engineering with common problems like memory, power, and compute,” Sutaria affirms.
Breakthroughs in embedded AI
One of the most notable technical advances in recent embedded systems—a point both engineers highlight—is real-time live translation. “The most exciting breakthrough is live translation,” Sutaria shares.
“Users speaking two different languages can talk to each other in their native language in real time. This is because of sophisticated AI, audio, and embedded engineering. Devices and robots can communicate with humans as they understand us more.” New architectures such as edge AI-powered speech translation devices demonstrate how hybrid cloud-edge designs are now practical for global use, achieving near-instant latencies and higher accuracy versus smartphone apps.
Continuous improvements in noise reduction, voice wake-up, and personalized audio experiences are made possible by advances in both device firmware and scalable AI infrastructure. The result is a proliferation of user experiences that were previously unthinkable in consumer electronics.
Learning across ecosystems
Innovation is rarely the product of isolated effort. Sutaria and Shah’s relationship is defined by the free exchange of ideas, even across company boundaries. “When you break an innovation down to the lowest level of design and coding principle people would be thrilled to see they all follow the same principles and style,” Sutaria notes.
Such principles transcend organizational differences, as engineers focus on system design, debugging methodologies, and best practices in code quality. “We understand how critical each other’s work is and what happens if it gets delayed,” Shah relates, highlighting how practical cross-pollination can sharpen both execution and design rigor for both device and cloud teams.
These exchanges are situated within broader shifts in collaboration. Regulatory actions, such as the EU’s requirement for interoperability between Apple and Google, actively encourage companies to remove artificial ecosystem barriers, transforming once-proprietary platforms into more open and interconnected systems. As a result, the collective knowledge base—and the speed of product advancement—rises for the entire industry.
Engineering for power and scale
At the core of modern embedded system design is the tension between increasing power and efficiency. “The biggest challenges are the same,” Sutaria stresses.
“Scaling down the cost of the product, increasing available compute, lowering power requirements, scalable software usable by millions, and defining and testing the product.” Shah’s approach to power optimization aligns with best practices in chip design for the IoT era, where dynamic voltage and frequency scaling, advanced ASIC architectures, and energy-aware software integration are essential to scaling up intelligent devices while containing energy cost.
For consumer electronics, these constraints become more acute as products shrink and user workloads intensify. AI wearables, for example, increasingly employ application processor plus coprocessor architectures or MCU-only designs, optimizing the tradeoff between battery life, compute power, and physical form factor. Both industries rely heavily on early validation and robust test automation—tenets drawn from large-scale infrastructure but ever more critical as device AI becomes foundational to consumer experience.
Impact at scale
Direct lines connect embedded engineering innovations to daily life. Sutaria points out, “Fifteen years back, we did not have any of this, and now we can see how it has enhanced people’s lives. AirPods are the world’s first OTC hearing aid to run ChatGPT running on Google Cloud. We think it has improved people’s lives; it has made people connected.” Shah echoes that the true impact of their work is made visible by its ubiquity—products, features, and protocols now woven into the routines of billions.
The evolution of major platforms—such as the Apple-Google collaboration on AI foundation models—is accelerating this transformation. The rise of privacy-focused on-device AI, secured via techniques like multi-layered encryption and hardware-level attestation frameworks, ensures that large-scale innovation serves practical needs while adhering to growing expectations for data security and user trust.
The next decade
The boundaries between embedded AI, audio technology, and power management continue to dissolve. “The field is going to evolve more and more embedded AI, audio would need more and more compute, and the wearables segment and robotics segment are going to test limits of what the system can do,” Sutaria notes. As next-generation architectures emerge, edge AI trends point to the adoption of specialized chips, federated learning, and privacy-preserving local processing in verticals from healthcare to automotive.
This convergence is fueling the development of products that can sense, react, and personalize in real time, blurring distinctions between local and remote intelligence, and requiring new models of joint validation and interoperability between device and infrastructure teams. Effective standards for IoT product development are increasingly being shared and embedded, turning the entire engineering workflow into an iterative, community-driven process.
Growth and advice
Professional growth, in Sutaria and Shah’s experience, is sustained by mutual understanding and support. “Since we both work in the same field, we understand challenges and critical portions of each other’s jobs. That helps us to grow in our company. Sometimes, late-night work and cross-country communication make a job difficult, but understanding each other makes it easier.”
Sutaria reflects on motivation: “Embedded devices are expensive to produce. It’s a huge investment. If you ever feel frustrated, just think of joy when it’s used by customers; customer satisfaction is what people should worry about. The rest will fall into place.”
As AI-driven embedded engineering becomes both more powerful and more distributed, the long-term lessons from their partnership are clear—success is built not just on technical insight but on relentless focus on the user, cross-disciplinary collaboration, and the willingness to operate across boundaries of company and domain.
Sutaria and Shah’s partnership reveals a defining trend in embedded engineering: technical innovation gains real, global traction only when expertise in device intelligence and infrastructure scalability is integrated. Their approach not only sets performance benchmarks across consumer and enterprise sectors but also highlights the importance of thoughtful engineering leadership in building technology that is both impactful and enduring.
:::tip
This story was distributed as a release by Jon Stojan under HackerNoon’s Business Blogging Program.
:::
