Turning data platforms intodurable business capability.

I build GenAI-powered analytics systems that turn complex data into self-serve insights and decisions.

Data, AI & Analytics Leader · Los Angeles, CA

About

I'm an enterprise data & AI leader who turns complex technology into trusted, real-time decisions — architecting lakehouse and agentic AI platforms that let Fortune-scale leaders act with confidence, not guesswork, at every level of the business.

Trusted decisions, not just dashboards

I build Databricks Genie-powered conversational analytics (Nalytics IQ) so leaders get plain-English answers they can act on immediately — replacing static reports with real-time, trustworthy intelligence.

Agentic AI that moves the business

I design and ship LangChain/LangGraph ReAct agents that automate multi-step business workflows end to end — from procurement decisions to data-quality checks — turning AI from a demo into measurable operating leverage.

Governance that earns executive trust

I extend enterprise-grade governance — security, lineage, auditability — into AI agent access and cost control, so leadership can scale AI adoption without scaling risk.

Technology in service of business outcomes

Every architecture decision I make is judged by the business impact it unlocks — faster decisions, lower cost to serve, and new revenue opportunities — not just technical elegance.

Work

01Product Studio

Nalytics IQ

Ask a question. Get the analysis.

An AI analytics agent on Databricks Genie that turns a plain-English business question into governed SQL, charts, and a written executive analysis — without an analyst in the loop.

One run, condensed to four moments: question, plan, comparison, exported report.

Questions it answers (sample)

List customers with a 20% or more decline purchase value in 2025 compared to 2024.
Identify clients completely inactive for the last 6 months.
What are the top 10 performing products in 2025, and how are they performing in 2026?
Provide a write-up on the journey of active brands.

Problem

Every new business question meant a ticket, a queue, and a bespoke report. Dashboards answered the questions someone anticipated last quarter, not the one a leader has right now.

Approach

A curated gold-layer semantic model the agent can reason over, governed end to end by Unity Catalog. A research agent plans the analysis as multiple steps, runs them, and writes up what it found.

Outcome

One question produces a four-step analysis plan, the SQL behind each step, comparison charts, narrative analysis, CSV download, and a multi-page PDF report — in seconds, not the days a reporting ticket used to take.

02Product Studio

Nexa

Try me

Universal analytics experimentation. Any data. Any model. One conversation.

A personal product that provides businesses with a model-agnostic natural-language analytics layer, works across data platforms, and turns plain-English questions into SQL, result tables, and auto-generated charts.

A product walkthrough: ask, plan, analyze, and explore business insights.

Features

Any data source: Databricks, Snowflake, MySQL, Google BigQuery, CSV, and other.
Flexible LLM integration: connect OpenAI, Anthropic, or Gemini with an API key, or use an existing Claude or GitHub Copilot login.
Schema browser: load tables, select tables, describe them in natural language.
Multi-line natural-language question box for richer business asks.
Results insight: surface the hidden truth behind the data, supported by a table and an auto-generated bar, line, or pie chart.
Read-only guardrail to prevent destructive SQL.
Zero- or low-cost self-service GenAI-based BI for businesses.

Positioning

Any Data → Intelligence Layer → Any LLM → Natural Language → Trusted Insights. Nexa is designed as a model-flexible intelligence layer between enterprise data platforms and the business users who need answers.

Approach

Users connect a data source, provide an OpenAI, Anthropic, or Gemini key, browse schemas, select safe tables, and ask a multi-line natural-language question. Nexa generates read-only SQL, runs it, and renders both the result set and an appropriate chart.

Outcome

The product demonstrates how governed natural-language analytics can work without locking users into one warehouse or one model provider — making trusted insights portable across data and AI stacks.

03Product Studio

SpeakIQ

Try me

Real-time AI call assistant.

An experimental product exploring multimodal AI, agentic workflows, and human-AI interaction.

SpeakIQ live transcript and suggested-response panels during an assessment-mode session

Live transcript on the left, streamed suggested responses with quantified talking points on the right.

Features

Live transcript: real-time speech-to-text with interim results through the Web Speech API — free, in the browser.
System or tab audio capture: transcribe the other party on a call, not just the microphone.
Speaker separation: label conversation turns as You and Them when both sources are captured.
Suggested responses: AI drafts a natural, first-person reply as the conversation develops, or on demand with Answer now.
Mode presets: Assessment, Sales, Meeting, Support, and General Q&A.
Your context: paste or attach a resume, role, or product brief (PDF, Word, or text) to tailor suggestions.
Flexible LLM integration: sign in with Claude or GitHub Copilot, or bring an OpenAI, Anthropic, or Gemini key.
Export: download the transcript or transcript plus suggestions as PDF, text, or Markdown.

Problem

Live conversations move quickly, making it hard to retain every detail, respond clearly, and keep relevant context at hand — especially when the stakes are high.

Approach

SpeakIQ transcribes the microphone in the browser for free, optionally captures system or tab audio for the other party, separates speakers as You and Them, and streams an AI-drafted first-person reply as the conversation develops.

Outcome

A flexible conversation companion that gives users a live transcript, mode-aware suggestions, and exportable session notes when they need to respond with clarity and confidence.

Selected highlights

Retail Intelligence — AI-Powered Growth

Created an intelligent retail analytics experience that turns complex customer and business data into self-service insights. Empowered teams to discover growth opportunities, expand customer reach, and identify renewal risks earlier — driving 15% revenue growth, 10% customer-base growth, and 8% higher contract renewal rates.

AI AnalyticsGen AICustomer AnalyticsPredictive Insights

B2B ReAct Agent — LangChain & LangGraph

Designed a tool-use ReAct agentic AI system for B2B e-commerce that understands buyer/supplier queries, searches catalogs, compares vendor pricing, and reasons through multi-step workflows using LangGraph state machines.

LangChainLangGraphAgentic AIPython

Enterprise Databricks Lakehouse Platform

Designed and implemented a medallion architecture (bronze/silver/gold) on Databricks Lakehouse from the ground up, establishing modeling standards and Delta Lake best practices for a B2B marketplace serving 150K+ customers.

DatabricksDelta LakeAzureData Modeling

Unity Catalog Governance Framework

Implemented Unity Catalog namespace hierarchy for multi-domain data organization, enforcing row/column-level security and building comprehensive data lineage, dictionaries, and ERD documentation — the same governance discipline now extending to AI agent access and auditability.

Unity CatalogData GovernanceAI Governance

$20M+ Operational Impact via Data Optimization

Diagnosed and resolved Spark performance issues (data skew, small files, inefficient joins) and led SCD pattern implementation and pipeline modernization, contributing to $20M+ in measurable operational impact.

SparkPerformance TuningETL/ELT

Retail BI Automation — 1,400+ Locations

Automated analytics and reporting for 1,400+ retail locations at Sears Holdings, saving 120+ hours per week in manual reporting effort while partnering with business leaders on KPIs.

BI AutomationReportingRetail Analytics

Industry Recognition & Speaking

Ideas built at work, recognized beyond it.

A concise proof band for externally linked moments where embedded analytics, GenAI self-service, and data-platform strategy met broader market conversations.

01TDWI guide

Industry recognition

Recognized in the ecosystem for embedded AI-powered analytics

A ThoughtSpot-promoted TDWI guide frames best practices for embedding AI-powered analytics with a data lakehouse, emphasizing faster time to value, monetizable data experiences, and GenAI-powered self-service analytics.

Read report

02Data + AI Summit 2023

Conference stage · Speaker

Featured perspective on self-service adoption with embedded analytics

As a speaker at the Data + AI Summit, I shared perspective on enterprise analytics adoption and embedded, GenAI-powered self-service experiences in an industry event setting.

Watch session LinkedIn context

Journey

Expertise

Data & Analytics Leadership

Data StrategyBI StrategyGlobal Team LeadershipExecutive CommunicationStakeholder Alignment

AI & Analytics

Databricks GenieAgentic AILangGraphLangChainConversational AnalyticsBI tools (MicroStrategy, ThoughtSpot, ...)

Data Platforms

DatabricksDelta LakeSnowflakeMongoDBAzure

Architecture & Governance

LakehouseMedallion ArchitectureUnity CatalogData ModelingData Governance

Engineering

PythonSQLETL/ELTData QualityPerformance Engineering

Certifications

GitHub Copilot CertifiedAWS Solutions ArchitectAzure Data EngineerDatabricks Data EngineerAI/ML EngineerThoughtSpot Data ExpertMicroStrategy Architect

Education

MS, Management Information Systems

University of Illinois, USA · 2011

BS, Engineering

NIT Patna, India · 2007

Experience

Jan 2022 — Present

Data, AI & Analytics · Forge (ex EmeraldX) — Los Angeles, CA

Leading enterprise Data & AI strategy supporting 150K+ customers
Architected Databricks Lakehouse from the ground up
Led migration from legacy BI platform to modern analytics architecture
Built Nalytics IQ, a GenAI conversational analytics capability
Established Unity Catalog governance across data and AI workloads
Leading a global team of 8+ engineers and architects

Apr 2018 — Dec 2021

Manager, Business Intelligence · Client: NBCUniversal | InfoCepts LLC — Los Angeles, CA

Led BI and Insights initiatives supporting digital media platforms and monetization strategies. Managed 20+ BI engineers and analysts across global teams, overseeing the end-to-end analytics lifecycle from ingestion to executive communication.

Jul 2017 — Apr 2018

Senior Software Engineer (Business Intelligence) · Client: Centene | Mastech Digital — St. Louis, MO

Optimized the enterprise BI ecosystem, reducing reporting artifacts by 30,000+ objects and improving platform performance while strengthening data governance and metric consistency.

Jun 2015 — Jul 2017

Architect / Senior BI Engineer · Sears Holdings — Chicago, IL

Designed and maintained large-scale BI dashboards for supply chain and merchandising performance. Automated analytics for 1,400+ retail locations, saving 120+ hours per week in manual reporting effort.

2007 — 2015

Earlier Roles · HCA, Citi Bank, Verizon Wireless, IL Dept. of Public Health, HPCL, Tata Motors

Progressively advanced through BI Engineering, Data Architecture, and Enterprise Analytics roles (Technical Specialist/Leader, Programmer Analyst, Assistant Data Manager, Project Manager), building strong foundations in SQL, statistical analysis, and large-scale reporting systems.

Let's connect

Start a conversation.

Interested in building data + AI products that create measurable business impact? Let's talk.

LinkedIn GitHub