Data & AI Leader | Generative AI Architect | Innovation Catalyst
AI/ML leader with 22+ years from automotive manufacturing to enterprise GenAI. Building production multi-agent systems, semantic search engines, and RAG architectures at scale. Track record of award-winning solutions across manufacturing, aerospace, telecom, and BFSI with measurable business impact. Systems thinker bridging strategy, architecture, and delivery.
I lead AI engineering at Korn Ferry Digital, where my teams build enterprise-scale generative AI capabilities for the Talent Suite platform. My current work spans LLM applications, RAG systems, multi-agent workflows, MCP-based AI platform engineering, ML-ranked hybrid search, and the MLOps/LLMOps practices needed to operate these systems responsibly.
My career has an arc from manufacturing and quality engineering and data driven problem solving, to data science and AI, to Generative AI and leadership. My experience spans automotive, aerospace, fluid machinery, off-road equipment, and in a consultative setting, I've worked in energy and renewable telecommunications, and BFSI domains. The last decade of my career has been all about building Data and AI based systems.
Outside of work, I build through AI Explorations: open-source libraries, research tools, and working demos across search, agentic AI, AI for mathematics, topological data analysis, and dynamical systems. These projects are continuations of things I've come to learn and work with in some small capacity in the past. They are a way to pressure-test ideas I have about AI systems, search, data analysis, and the like through building tools and systems around sub-problems as open-source projects.
Leading AI engineering teams and platform strategy for enterprise-scale GenAI systems, with focus on innovation, architectural excellence and delivery of AI systems.
Architecting LLM applications, RAG systems, agentic workflows, MCP-based tools, and hybrid search systems that combine lexical, vector, and ranking signals.
Building production ML and LLMOps practices: model and data versioning, observability, evaluations, bias and fairness reviews, deployment discipline, and platform reliability.
Managing data lakes, building and deploying event-driven systems, and large-scale applications with attention to good architecture, clear separation of concerns, cost management, reliability and maintainability, operability, and organizational adoption.
Deep domain expertise across automotive, aerospace, industrial products, off-road equipment, manufacturing engineering, quality management, Lean Six Sigma, and data-driven problem solving.
Building open-source systems that explore search, multi-agent coordination, mathematical parsing, topological analysis, and dynamical systems as practical engineering artifacts.
A Pythonic multi-agent framework built around peer agents, structured spores, memory, tools, and observability. Praval explores agent ecosystems without a central orchestrator.
An agentic research assistant built on Praval that helps discover, index, synthesize, and discuss research papers. It treats research as synthesis, not just retrieval.
A Rust/Python search engine for lexical, vector, and hybrid retrieval. Vajra powers both the documentation search demo and the newer 3D shape search demo.
A neural-symbolic math system that learns to parse mathematical notation into structure, then delegates computation to deterministic symbolic tools instead of training arithmetic itself.
A topology-driven interpretability project for measuring the shape of neural representations using persistent homology, activation geometry, and model-comparison pipelines.
A dynamical-systems analysis toolkit for studying neural network training trajectories, perturbation sensitivity, Lyapunov exponents, and architecture-level stability.
Feel free to reach out for discussions about AI research, technical collaborations, or questions about any of my projects.