I help startups and growing businesses analyse their operations, redesign their processes, and deploy AI — in that exact order. No shortcuts, no hype, just systems that work.
Each stage earns the right to the next.
Before touching a tool or workflow, I map what the business needs to achieve and where it leaks value. Every recommendation is grounded in commercial reality — not what's technically interesting.
Good automation on a bad process gives you faster errors. Using Lean and BPR principles, I redesign workflows from first principles before any technology is introduced.
Only once business logic is clear and the process is sound do I bring in GenAI, automation, and analytics — as amplifiers of a proven system that pays for itself.
Map your value chain, find where time and money leak, and produce structured improvement roadmaps grounded in your commercial reality.
End-to-end workflow redesign using Lean methodology — turning chaotic operations into scalable, automatable systems.
Practical GenAI and automation deployments that go into production, not into a slide deck. Built on your actual processes.
Vendor-agnostic guidance on where to invest in technology, in what order, and how to roll it out without disrupting the business.
Statistical models and decision-science frameworks that turn operational data into strategic insight.
C-suite advisory for businesses in transition — operating model redesign, CRM implementation, change management.
Not hypotheticals — systems built, shipped, and measured.
200+ daily field appointment tickets handled automatically. WhatsApp first, an LLM voice call only on exceptions.
Read case studyOCR pipeline extracts, validates, and fuzzy-matches 30 supplier invoices a day — catching a 4% error rate the team was missing.
Read case studyWhisper transcribes field-rep voice notes on-premise. An open-source LLM structures them into CRM updates. Capture went from 40% to 100%.
Read case studyRAG pipeline behind a phone + PIN gate. Investors query revenue, projections, cohorts, churn — in plain English, in seconds.
Read case studyA voice bot runs the first partner conversation end to end. It costs 25% more per call than a human — and was still the right decision.
Read case studyIts first five outputs were rejected at approval. The model was working perfectly — the documents underneath it were lying.
Read case studyConcept layouts built with deterministic geometry, not a generative model — because a drawing that feeds an investment decision has to be checkable.
Read case studyEstimating logic ported out of Excel with the original cell references preserved, so the client's own engineers can audit every figure.
Read case studyDetailed blueprints showing exactly what gets automated, how it works, and what outcomes to expect — before you spend a rupee.
Advanced Certificate in Data Science & Decision Science
MBA — Marketing
B.E. — Electronics & Power Engineering
Tell me what you're working on. I'll tell you honestly whether I can help.
Bengaluru, India