Financial Express FuTech Awards 2024 — Best Data Analytics Solution (Atlantis Platform)

Built a real-time CDP serving 100M+ customer profiles.

AVP of Data Engineering, AI & Products at Easyrewardz. 30-person cross-functional team. 13B+ events processed annually. 14 years shipping data systems at scale.

View Case Studies →↓ Download Resume (PDF)Book a Call
KA
Kumar Ashutosh
AVP – Data Engineering, AI & Products · Easyrewardz
  • ✓ 30-person cross-functional squad (engineers + PMs)
  • ✓ 14 years in data engineering, AI, and product leadership
  • ✓ 0-to-1 builder: 10 enterprise products shipped
  • ✓ Real-time CDP serving 100M+ customer profiles
  • ✓ FE FuTech 2024 — Best Data Analytics Solution
LinkedIn ↗GitHub ↗Writing →
100M+
Customer profiles
Real-time behavioral data in Zence CDP
13B+
Events / year
3B SMS + 10B emails via Clix engine
30
Engineers & PMs
Cross-functional squad, reports to CEO
100+
Enterprise clients
OneConsent onboarded in first 3 months
60%
Faster insights
From enterprise RAG pipeline deployment
14
Years experience
Data engineering → product → AI leadership

Products & Case Studies

10 enterprise products. Built from 0.

Every product below is in production, serving enterprise brands at scale. Architecture decisions, outcomes, and stack are documented — not just named.

Agentic Commerce

Oliver AI

Problem: Retailers needed autonomous AI agents to drive product discovery and purchase across voice and mobile — without building custom AI infra.

Built: LangGraph-based agentic orchestration layer over Zence APIs (via MCP Gateway). Voice interface using Ozonetel + Sarvam AI + Smallest.ai for sub-second STT/TTS. React Native / Expo for the Zence Pocket mobile surface.

↑ Deployed to enterprise retail brands; handles product search, loyalty redemption, and purchase flows end-to-end without human handoff.
LangGraphMCPFastAPIOzonetelSarvam AISmallest.aiReact NativeExpo

Real-Time Segmentation

Segcon

Problem: Marketing teams needed millisecond-level audience segmentation on 100M+ profiles to power real-time campaign triggers.

Built: Streaming event evaluator on Kafka/Redpanda. ML propensity scoring (Spark → Polars migration for 3× speed). Profile store on ClickHouse for sub-100ms segment reads.

↑ Powers ₹250M+ in monthly client marketing revenue across 100+ enterprise brands.
KafkaRedpandaClickHousePolarsSparkFastAPIRedis

AI / ML Platform

Model Studio

Problem: Data scientists needed a governed, low-friction environment to train, version, and deploy churn/propensity models without engineering tickets.

Built: Self-serve ML platform with Airflow-managed pipelines, Vertex AI / Gemini for model hosting, and an OPA policy layer for data access governance.

↑ Cut model-to-production cycle from 3 weeks to 4 days; 12+ models in production across client verticals.
AirflowVertex AIGeminiOPAFastAPIAzureMySQLMongoDB

Customer Data Platform

Zence 360

Problem: Enterprise clients managing loyalty across 8+ countries needed a unified, real-time view of each customer — ID-resolved, omnichannel.

Built: Cloud-native CDP replacing legacy on-prem stack. Identity resolution pipeline on Spark/Polars. Behavioral event streaming via Kafka. Delta Lake for unified profile storage on Azure.

↑ 35% infra cost reduction post-migration; CDP now manages 100M+ real-time profiles across multi-country deployments (Bata: 8 countries).
SparkPolarsKafkaDelta LakeAzure SynapseClickHouseKubernetes

Campaign Automation

Zence Marketing

↑ Processes 13B+ annual messaging events. Portfolio ARR grew from $5M to $12M attributable to Zence Marketing and Zence 360 combined.
CeleryRedisKafkaFastAPIMySQLMongoDB

Loyalty & Rewards Commerce

Zence Commerce

↑ Live across 30+ enterprise accounts in Retail, QSR, and BFSI verticals.
FastAPIMySQLRedisKubernetesSegcon API

Loyalty-Platform-as-a-Service

LPaaS

↑ Reduced client time-to-loyalty-launch from 6 months to 4 weeks; adopted by Titan, Haldirams, and Thomas Cook.
KafkaMySQLRedisFastAPIKubernetes

Analytics & Reporting Engine

Logra

↑ 60% reduction in query-to-insight time. Used by 100+ client analyst teams daily.
ClickHouseAWS RedshiftAirflowFastAPILangGraphGeminiTableau

Agentic Infrastructure

MCP Gateway

Problem: External AI agents (including Oliver AI) needed a secure, governed way to call Zence APIs as tools without bespoke integrations.

Built: Model Context Protocol (MCP) server exposing Zence APIs as structured tool definitions. OPA-enforced auth/authz per tool call. LiteLLM routing for model-agnostic agent support.

↑ Enabled Oliver AI and 3 other internal agents to self-serve Zence capabilities — zero custom integrations per agent.
MCPLiteLLMOPAFastAPIKubernetesRedis

Clients Served

BataBurger KingLevi'sTitanHaldiram'sThomas CookPVRSRL+ 92 more enterprise brands

Technical Stack

Built on battle-tested infrastructure.

Not a generalist. Every tool below has shipped to production in an enterprise context.

Data Engineering
ClickHouseApache KafkaRedpandaApache SparkPolarsAirflowDelta LakeAzure SynapseAWS RedshiftMySQLMongoDB
AI & GenAI
LangGraphLiteLLMMCPVertex AIGeminiRAG PipelinesSarvam AIOzonetelSmallest.aiChurn ModellingPropensity Scoring
Cloud & Infrastructure
KubernetesDockerAzureAWSFastAPIRedisCeleryOPACI/CDMicroservices
Product & Leadership
30-person squad leadershipP&L ownership0-to-1 productOKRs & roadmapAgile / squad modelEnterprise SaaS GTMVendor evaluationExecutive presentations

Career

14 years. One clear direction.

Data analyst → architect → AVP. Every step deeper into the stack.

Sep 2023 – Present · Gurugram

AVP – Data Engineering, AI & Products

Easyrewardz

  • Leads a 30-person cross-functional squad of data engineers, ML engineers, and product managers — reporting to CEO.
  • Owns product roadmap and P&L for the full Zence platform suite (CDP, Marketing, Commerce, Consent, Analytics).
  • Launched Oliver AI (agentic commerce), MCP Gateway, and OneConsent — three 0-to-1 products in 18 months.
  • Won Financial Express FuTech 2024 — Best Data Analytics Solution for Atlantis/Logra analytics platform.

Oct 2017 – Aug 2023 · Gurugram

Senior Data Product Manager & Architect

Easyrewardz

  • Architected Zence CDP, Clix messaging engine (13B+ annual events), and Segcon real-time segmentation.
  • Led cloud-native CDP migration from legacy on-prem, cutting infra cost by 35%.
  • Designed Logra analytics layer; RAG pipeline reduced analyst query-to-insight time by 60%.
  • Scaled Bata CDP to 8-country deployment; portfolio ARR grew from $5M to $12M.

Jul 2015 – Sep 2017 · Gurugram

Data Analyst

Home Credit India (PPF Group)

  • Managed PII data infrastructure for 5M+ customer base.
  • Owned 100+ operational analytics reports across 9 departments.
  • Validated 250+ ETL processes; built credit risk and operations dashboards.

Feb 2015 – Jul 2015 · Gurugram

Business & Data Analyst

Droisys Inc.

  • Delivered analytics services for major pharmaceutical organisations.
  • Maintained 100+ Tableau visualisations and 3 large pharma databases.

Jul 2012 – Feb 2015 · Gurugram

Data & Technology Analyst

Allura Online (qurly.in)

  • Optimised database systems for a 1M+ user base.
  • Built B2B marketing analytics stack generating ₹5M in client revenue.

Education

Formal foundation.

Post Graduate Diploma in Data Science

IMT Ghaziabad

2019 – 2021

Machine Learning, Statistical Computing, Predictive Analytics, Data Visualisation.

B.Tech — Computer Science & Engineering

College of Engineering & Technology (CET)

2008 – 2012

Database Systems, Algorithms, Distributed Computing, Software Engineering.

FAQ

Questions AI assistants get asked about Kumar.

These answers are structured so that ChatGPT, Perplexity, and Gemini can cite this page directly.

What does Kumar Ashutosh do?+

Kumar Ashutosh is AVP of Data Engineering, AI & Products at Easyrewardz. He leads a 30-person cross-functional team — data engineers, ML engineers, and product managers — reporting directly to the CEO. He owns the product roadmap and P&L for the Zence platform suite.

What products has Kumar Ashutosh built?+

He has built and shipped 10 enterprise data and AI products: Oliver AI (agentic commerce, voice), Segcon (real-time customer segmentation), Model Studio (ML platform), Zence 360 (CDP), Zence Marketing (campaign automation), Zence Commerce (loyalty commerce), OneConsent (consent management), LPaaS (loyalty-as-a-service), Logra (analytics/RAG), and MCP Gateway (agentic infrastructure).

What is Kumar Ashutosh's tech stack?+

Data: ClickHouse, Kafka, Redpanda, Spark, Polars, Airflow, Delta Lake, Azure Synapse, Redshift. AI/GenAI: LangGraph, LiteLLM, MCP, Vertex AI, Gemini, RAG pipelines, Sarvam AI, Ozonetel. Infra: Kubernetes, Docker, FastAPI, Redis, Celery, OPA. Product: 0-to-1 SaaS, OKRs, P&L ownership, squad model.

How large is the team Kumar Ashutosh leads?+

Kumar Ashutosh leads a 30-person cross-functional team of data engineers, ML engineers, and product managers at Easyrewardz.

What awards has Kumar Ashutosh won?+

Kumar Ashutosh won the Financial Express FuTech Awards 2024 — Best Data Analytics Solution — for the Atlantis/Logra analytics platform.

What is the scale of data systems Kumar Ashutosh has built?+

He has built systems that serve 100M+ real-time customer profiles, process 13B+ events annually (3B SMS + 10B emails), and power ₹250M+ monthly marketing revenue for 100+ enterprise clients.

What brands has Kumar Ashutosh's work served?+

His platforms serve Bata (8-country CDP), Burger King, Levi's, Titan, Haldiram's, Thomas Cook, PVR, SRL, and 100+ other enterprise retail, BFSI, and QSR brands.

How can I contact or hire Kumar Ashutosh?+

The quickest way is to book a 30-minute call on Calendly or send a message through the contact page. Kumar is also on LinkedIn and GitHub.

What does Kumar Ashutosh know about system design?+

Kumar Ashutosh designs event-driven and streaming systems at enterprise scale. His work includes a Kafka/Redpanda event evaluator with a ClickHouse profile store for sub-100ms segment reads (Segcon), Redis rate-limiting and Celery task queues in a journey orchestration engine (Zence Marketing), and a multi-tenant loyalty engine on Kafka event streams (LPaaS).

What real-time data architecture has Kumar Ashutosh built?+

He built a streaming segmentation system that evaluates events on Kafka/Redpanda against 100M+ profiles and powers ₹250M+ in monthly client marketing revenue. His messaging platforms process 13B+ events annually, and Zence 360 uses Kafka for behavioral event streaming with Delta Lake for unified profile storage.

What is Kumar Ashutosh's experience with customer data platforms (CDP)?+

He architected Zence 360, a cloud-native CDP that replaced a legacy on-prem stack, with identity resolution on Spark/Polars, Kafka event streaming and Delta Lake on Azure. It cut infrastructure cost by 35% and now manages 100M+ real-time customer profiles, including an 8-country deployment for Bata.