SELECT leader FROM data_and_ai WHERE years > 18 AND ships = TRUE;
I build the teams, platforms and AI-assisted engineering practices that turn enterprise data into decisions, and I run the delivery that gets them into production.
-- operating thesis
Eighteen years have taught me that models are the easy part. The hard part is the platform underneath, the team around it, and the commercial model that funds it. I lead all three.
Multi-agent systems that audit, reason and act, built with deterministic routing, hard invariants and human approval gates so they are safe to run in an enterprise.
LangGraph · Claude Agent SDK · MCP · Ollama · Qdrant
From SQL Server and SSAS to Azure, Snowflake and dbt. Semantic layers that cut BI cost, governed pipelines, and metadata that makes change safe.
Snowflake · dbt · ADF · Synapse · Purview · Tableau · Power BI
Pre-sales architecture, SOWs, WBS and estimation, and multi-team account delivery for a Fortune-500 high-tech client. I design services that are repeatable and profitable.
Pre-sales · SOW & pricing · Multi-team delivery · QBRs
-- selected work
Open-source projects where I prototype the patterns my teams later take to production.
Turns a screenshot or PDF of any dashboard or report into a client-safe replica: every sensitive name, ID and number is masked on your device and replaced in place with synthetic text, so the charts and layout look exactly like the original. An independent six-pass re-scan checks the masking, and a safe-to-share receipt with the image's SHA-256 proves what was hidden. Nothing is uploaded and no AI service is called.
WHY IT MATTERSPre-sales, BI and product teams can show a client's dashboard, including real customer verbatims with the people removed, without exposing a single real number. Free, no sign-up, no backend and no stored data. Live on Cloudflare Pages, with the Chrome and Edge extensions in store review.
A three-agent adversarial code and FinOps auditor that runs fully offline on a laptop CPU, on a 398 MB local model. No API keys, no cloud cost, no data leaving the machine.
WHY IT MATTERSShows how regulated enterprises can get AI review without sending code or cost data to a third party.
Diffs Tableau workbook revisions and produces HTML/PDF reports with structural changes, dependency impact, lineage and a risk score, answering "what changed in the last publish, and is it safe?"
A multi-agent travel planner: flight, hotel, itinerary and final-answer agents chained in LangGraph, using AviationStack, Tavily and Groq, with conversation state checkpointed to PostgreSQL.
Finds costly warehouse query patterns with Qdrant named vectors and enforces hard invariants such as "never downscale if the historical match breached its SLA" before recommending a fix.
Replace this card with a new repo: title, one-line problem, stack chips and link.
-- career lineage
Each role added a layer: engineering, then architecture, then people, then commercial ownership.
Ascendion · Bengaluru
Lead a 40-person, multi-team delivery organisation for a Fortune-500 high-tech client across data engineering, BI and AI-assisted engineering, with close to $3M in projected annual revenue. Own pre-sales architecture, SOWs, estimation and account growth.
Absolutdata Analytics, an Infogain company · Bengaluru
Managed data science and BI programmes on AWS, Microsoft Power Platform and Azure DevOps as a customer-success lead, and contributed to the company's diversity and inclusion work as it expanded internationally.
Nous Infosystems · Bengaluru
Ran cloud data and BI projects end to end and set technical roadmaps for C-level and engineering audiences.
Nous Infosystems · Bengaluru
Designed data warehouse and BI solutions on the Microsoft stack and Azure.
Cerner India Health Services · Bengaluru
Led Microsoft BI delivery for healthcare analytics.
Siemens Technology and Services
Built data warehouses, SSIS/SSAS models and enterprise reporting.
ProSharpe Wealth Management
Where it started: applications, databases and a lasting love of SQL.
-- capabilities
-- credentials
-- writing
I've written about data and AI since 2018. Recent pieces:
Turning Tableau revision history into a readable, risk-scored change report.
Moving logic left into Snowflake to replace billion-row extracts with live connections.
Near real-time change data capture without hand-built pipelines.
Scaling data ownership across domains in large enterprises.
Lineage and discovery without querying the repository database.
Replace with the title, date, one-line summary and link.
-- open to
Senior Director and Head of AI / Data roles where I can build teams, platforms and a delivery engine together. Based in India, working with global teams.