Sudeep George believes that effective technology leadership cannot happen from a distance. Across a career spanning more than two decades in computer vision, custom silicon, computational imaging, and sensor fusion, his work has been guided by a consistent principle that strategy must be grounded in how complex systems are actually built, operated, and improved. Now, as Chief Technology Officer of iMerit Technology, he brings his engineering perspective to the development of multimodal, data-centric solutions used to train, evaluate, and refine advanced AI models.

Sudeep became CTO in December 2024 after serving as vice president of engineering. His earlier career included system software development at Samsung Research India and building computer vision platforms around proprietary ASICs at Sarnoff Corporation. He later co-founded Tonbo Imaging, where he led engineering for computational imaging products built around edge computing, sensor fusion, and custom hardware. Across those roles, he has created and scaled multidisciplinary teams working through complex hardware and software challenges.

In this conversation with The Consulting Report, Sudeep breaks down iMerit's approach to data annotation and Generative AI services, how the firm operationalizes expert judgment through its proprietary AngoHub platform, and how the shift toward domain-specific AI is fundamentally changing client demands.

This interview has been edited for length and clarity.

The Consulting Report: Can you provide an overview of iMerit’s industry and functional areas of specialization?

Sudeep George: iMerit specializes in providing expert data for AI systems operating in mission-critical environments. We focus on sectors where accuracy, safety, trust, and domain expertise are essential, including healthcare, autonomous mobility, agriculture, government, and generative AI. Our clients include three of the "Big 7" generative AI companies, eight of the top autonomous vehicle companies, three large U.S. government agencies, and two of the top three cloud providers. 

Across these sectors, the common challenge is that AI systems are moving from controlled development environments into real-world decision-making. That shift requires more than large volumes of labeled data; it requires expert judgment, structured workflows, rigorous quality systems, and the ability to identify ambiguity, edge cases, and model failure modes.

iMerit combines its technology platform, AngoHub, with specialist human expertise to support critical parts of the AI development lifecycle. This includes data annotation, model evaluation, validation, quality assurance, failure-mode analysis, and expert feedback loops. In areas such as autonomous mobility and physical AI, our work spans complex multimodal data, including images, video, LiDAR, point clouds, and sensor-fusion workflows. In generative AI and healthcare, we support expert-led evaluation, reasoning assessment, domain-specific data creation, and quality review.

Our mission is to help clients build AI systems that perform reliably not just in benchmark conditions, but in the complex and ambiguous conditions they face in production.

“We do not view human expertise as a commodity.”

The Consulting Report: How does your firm differentiate itself from competitors?

Sudeep George: We differentiate ourselves through the combination of technology, domain expertise, and a unique approach to developing and operationalizing the human expertise that powers modern AI systems.

A top differentiator is our depth in complex AI data. Many firms can support basic annotation at scale. iMerit’s strength is in high-complexity, high-judgment workflows where output quality depends on expertise, context, and process design. In autonomous mobility and robotics, for example, our teams use AngoHub to work across 2D and 3D data, LiDAR, point clouds, video, object tracking, sensor fusion, and edge-case analysis. In healthcare and generative AI, we support workflows requiring domain specialists, structured model evaluation, and rigorous review.

AngoHub is central to how we operationalize this expertise. The platform enables configurable workflows, quality gates, ontology management, advanced labeling, reviewer feedback, and secure delivery processes. This allows us to design repeatable, measurable workflows rather than relying on ad hoc human effort alone. For clients operating in regulated, safety-sensitive, or high-trust environments, that level of control and traceability is critical.

Another strong differentiator is the iMerit Scholars program, a handpicked global network of subject-matter experts selected for their domain knowledge, credentials, and broader cognitive abilities. These experts help teach, challenge, evaluate, and improve AI models.

As enterprises move from general-purpose AI to domain-specific applications in high-stakes sectors, success increasingly depends on expert judgment rather than scale alone. iMerit Scholars include physicians, mathematicians, scientists, linguists, and other specialists who provide the deep reasoning, contextual understanding, and human oversight required to train and evaluate advanced AI systems and edge cases. 

Ultimately, our differentiation is that we do not view human expertise as a commodity. We combine expert judgment, workflow technology, quality systems, and long-term operational accountability to help clients improve AI performance in the areas where mistakes are most costly.

The Consulting Report: Are there one or two major client projects that demonstrate your firm’s capabilities?

Sudeep George: One example is iMerit’s work with Digital Green, a nonprofit developing an AI-powered assistant that helps smallholder farmers access agricultural advice through voice, text, and image-based queries. As the platform scaled, Digital Green faced a critical challenge: existing speech-to-text models struggled with low-resource languages, regional dialects, and agriculture-specific terminology, making accurate voice recognition difficult. iMerit partnered with Digital Green to build a scalable human-in-the-loop transcription pipeline using AngoHub, combining native-language expertise, multi-layer quality assurance, and structured workflows to create high-quality training and evaluation datasets. 

The collaboration produced more than 100 hours of high-quality agricultural voice data, established a benchmarking framework for evaluating speech models in real-world conditions, identified critical model failure modes, and accelerated model fine-tuning efforts. These improvements support the assistant’s ability to deliver more accurate, context-aware recommendations to farmers at scale, helping the platform serve millions of agricultural queries across multiple languages and regions. 

The Consulting Report: How would you describe your firm’s culture?

Sudeep George: iMerit’s culture is built around the belief that people and technology must evolve together. AI quality is not achieved by technology alone. It requires strong platforms, disciplined processes, domain expertise, and people who understand the real-world consequences of the work they are doing.

That belief has shaped how iMerit builds its workforce. We have invested in long-term career pathways for AI practitioners and subject-matter experts rather than relying only on transactional gig-work models. Our teams are trained not just to complete tasks, but to understand client objectives, quality expectations, edge cases, and the downstream use of the data or feedback they produce.

This philosophy is also reflected in initiatives such as the Scholars program, which fosters a global community of experts who actively contribute to the development of AI within their own fields. With this approach, iMerit has built a highly engaged workforce, reflected in a 91% retention rate among its experts and employees.

“As AI moves closer to real-world deployment, the need for trusted human expertise becomes more important, not less.”

The Consulting Report: Has the increasing prevalence of AI changed the types of client mandates you are hired for?

Sudeep George: Yes. As AI has matured, client demand has shifted significantly from large-scale data labeling toward higher-value services centered on expert knowledge, model evaluation, reasoning, and domain-specific customization. We are seeing a broader industry transition from prioritizing data quantity to prioritizing data quality and specialized expertise. Increasingly, enterprises and AI developers are seeking support for model fine-tuning, reinforcement learning, evaluation frameworks, chain-of-thought reasoning, multimodal AI systems, and production-grade AI deployments in regulated or high-stakes industries. 

A clear example is in autonomous mobility and robotics. Earlier AI programs were often focused on annotating images or sensor data to train perception models. Today, clients are asking us to help evaluate how AI systems behave in complex real-world environments, across multimodal scenarios that combine video, LiDAR, point clouds, audio, and language. Increasingly, the work involves identifying edge cases, building evaluation environments, testing how autonomous systems respond to unusual situations, and providing expert oversight to improve reliability and safety. 

As AI moves closer to real-world deployment, the need for trusted human expertise becomes more important, not less.

“Strategy is important, but it has to be grounded in how systems are actually built, operated, and improved.”

The Consulting Report: How do you like to spend your time outside of work?

Sudeep George: Outside work, I enjoy spending time with my family and staying connected to technology beyond the immediate demands of the job. I have always been curious about how systems work, whether that is software, hardware, cars, imaging systems, or new AI tools.

I also try to stay hands-on with technology. For me, effective technology leadership requires staying close to real engineering problems. Strategy is important, but it has to be grounded in how systems are actually built, operated, and improved. That curiosity continues to shape how I approach both work and life.