18+ years across engineering, AI/ML, and product leadership. Today I lead Agentbase, DB AI Chat, and Technographics at Demandbase — owning enterprise AI platforms with globally distributed teams across the US, India, and Russia.
I lead AI & Engineering at Demandbase, owning Agentbase (our enterprise multi-agent platform), DB AI Chat (a unified conversational layer over the product suite), and Technographics (TB-scale signal intelligence).
Before that, I scaled a data science platform at Embibe from ~100 to ~1M users/sec, served as VP Engineering at FundsCorner helping build an INR 90 Cr loan book, co-founded Gathr and 501 Labs, and started out at Tech Mahindra delivering enterprise systems for British Telecom.
I specialize in turning raw data into scalable AI products, orchestrating intelligent agents that think and act, and building 0→1 in high-stakes environments.
Metrics, traces, and logs were designed for deterministic services. Multi-agent systems are non-deterministic, multi-step, and self-correcting — which means we need new primitives for trust, evaluation, and debuggability. We're building them at Demandbase.
Real enterprise AI needs retrieval you can trust, a semantic control plane over your actual data, evaluation you can defend to legal, and latency budgets that survive production traffic. The interesting work starts after the model call.
Feature stores solved ML training. The next layer — a queryable, governed semantic map over thousands of tables and columns — is what lets agents act deterministically over messy, real-world data. This is the piece most enterprise AI stacks are still missing.
Progressed through Principal Data Scientist → Senior Data Science Manager → Director.
Lead AI and engineering across Agentbase, DB AI Chat, and Technographics. Own architecture and delivery for products contributing $10M+ in annual revenue; manage globally distributed teams across US, India, and Russia.
Led the AI Data Science Platform and Search. Re-architected search, shipped Search V2, and scaled the platform from ~100 to ~1M users/sec. Partnered with ex-Pinterest / Walmart Labs search leaders.
Built AI/ML & computer-vision for credit scoring, KYC, and cheque parsing. Designed microservices & serverless on AWS. Helped build a ~INR 90 Cr loan book and supported 20× business growth.
Co-founded Gathr (social discovery; national press coverage, full-stack ownership) and 501 Labs (product studio shipping iOS, Android, and web products). Built stacks, shipped products, grew users.
Delivered enterprise systems for British Telecommunications. Foundations in large-scale engineering & delivery discipline.
A system of connected AI agents for B2B go-to-market. Multi-agent orchestration, A2A protocols, MCP, a unified data foundation, and enterprise-grade governance & observability.
Unified conversational layer across the Demandbase suite. Multi-agent orchestration + semantic control plane over thousands of tables & columns delivering low-latency, deterministic, context-aware outputs.
Social discovery app for casual, short-notice activities. Built the entire stack — backend, web, native Android. Grew to 2,500+ users with national press coverage across The Hindu, Economic Times, Deccan Herald, and more.
TB-scale signal intelligence powering Demandbase's account intelligence. Own the platform end-to-end — pipelines, models, and a semantic layer that turns billions of raw signals into reliable, queryable B2B truth.
Re-architected the AI Data Science Platform & Search at Embibe (Reliance Jio). Shipped Search V2 and scaled from ~100 to ~1M users/sec, partnering with ex-Pinterest and Walmart Labs search leaders.
Built AI/ML and computer vision for credit scoring, KYC, and cheque parsing. Designed microservices & serverless architecture on AWS that helped grow a ~INR 90 Cr loan book and supported 20× business growth.
Always happy to connect with builders working on enterprise AI, agentic systems, and 0→1 products.