Q1. You’ve built a strong career leading AI-driven innovation across industries — could you share what inspired your journey into AI and how your focus has evolved over the years?

My journey into artificial intelligence started with a fascination for pattern recognition and decision systems during my engineering studies. Over the years, that curiosity evolved into a deep passion for developing intelligent solutions that not only solve problems but also anticipate them. Although I began with a focus on classical machine learning and predictive analytics, my interests have expanded to embrace generative AI, large language models, and agentic architectures. Today, I am dedicated to building enterprise-grade copilots, orchestrating sophisticated workflows with tools like LangChain and LangGraph, and ensuring that AI capabilities are tightly aligned with broader business objectives.

Q2. Many organizations are struggling to turn AI proof-of-concepts into scalable business solutions. What do you think are the most overlooked factors in that transition?

A common pitfall I see is the lack of robust evaluation frameworks and ongoing feedback mechanisms. Many proof-of-concepts showcase technical feasibility, but they often overlook critical aspects like robustness, explainability, and stakeholder alignment. From my experience, teams are more successful when they focus on retrieval accuracy, model observability, and embed human-in-the-loop processes into their designs. Strong data readiness, governance, and cross-functional ownership are also vital—especially when organizations are scaling AI across complex, multi-cloud, or regulated environments.

Q3. How is the shift toward AI governance and responsible AI practices shaping the way architects like you design and deploy models?

AI governance is no longer a mere recommendation—it's become a foundational necessity. With global regulations and ethical scrutiny intensifying, architects must build responsible AI principles into every stage of model development. That means actively addressing bias mitigation, transparency, and auditability from the beginning. Personally, I use tools like RAGAS for LLM evaluation and Azure AI Foundry’s governance modules to keep compliance and trust at the forefront. Today, governance directly influences technical choices, leading to a preference for modular, interpretable systems rather than opaque, monolithic ones.

Q4. With increasing use of multimodal models that combine text, vision, and speech — where do you see the most promising applications emerging?

Multimodal models are opening up transformative possibilities in fields like healthcare diagnostics, legal document review, and industrial automation. For instance, I recently led a proof of concept for a health insurer that combined large language models, optical character recognition, and clinical named entity recognition to automate underwriting—a clear example of how these technologies can work together. Bringing together text, vision, and speech is also revolutionizing assistive technology, smart manufacturing, and contextual copilots that understand not just language, but the broader environment.

Q5. As cloud  ecosystems mature, how do you see Azure’s AI stack evolving compared to AWS and Google Cloud in supporting enterprise-scale AI transformation?

Azure is quickly evolving into a highly developer-friendly and enterprise-ready AI ecosystem. Its seamless integration of Azure OpenAI, Prompt Flows, Semantic Kernel, and LangChain delivers a unified experience for building scalable generative AI solutions. In comparison, AWS offers strong modularity but can feel fragmented, while Google Cloud excels in research but is less cohesive for enterprise deployment. Azure stands out for balancing governance, orchestration, and business alignment. The introduction of AI Foundry and Copilot Studio, in particular, is proving to be a game-changer for enterprise adoption.

Q6. AI automation and copilots are transforming workflows across domains. In your view, what roles or processes are likely to be most disrupted in the next 2–3 years?

We’re on the cusp of major changes in knowledge work, customer support, and software development. AI copilots are already streamlining software development lifecycle processes, documentation, and data analysis. In the next two to three years, roles that involve repetitive decision-making, compliance reviews, and routine reporting will be fundamentally redefined. Crucially, the goal isn’t to replace professionals, but to augment their capabilities—freeing them to focus more on strategy and creativity while AI takes care of repetitive tasks.

Q7. From an  investor and innovation standpoint, which areas of AI — whether infrastructure, applied AI, or tooling — do you believe will drive the next major wave of growth?

Three areas stand out to me:

• AI Infrastructure: Vector DBs, orchestration frameworks (LangGraph, Semantic Kernel), and scalable evaluation tools.
• Applied AI: Domain-specific copilots in healthcare, finance, and legal.
• Tooling: Low-code/no-code platforms, agentic AI, and governance-first design systems.

Investor interest is shifting from flashy demos to platform-level resilience, compliance, and ROI. The next wave will be led by interoperable, secure, and human-centric AI systems.