Headline

Data Engineer @ Zepto | Real-Time Streaming & Batch Pipelines | Kafka · Flink · Spark · Databricks · AWS | Ex-Nielsen · Ex-Tredence |

About

I build data systems that businesses can't afford to have go down. Over 4 years, I've engineered large-scale data platforms across Quick Commerce, Audience Measurement and CPG domain— powering supply chain planning, real-time inventory intelligence, and operational decision-making at Zepto, Nielsen, and Tredence. At Zepto today, I own the Planning data platform driving automated replenishment across 1200+ dark stores, every SKU, every minute. What I've shipped: Architected a real-time Inventory Tracking System — Kafka event backbone → Flink stateful streaming → ClickHouse OLAP — replacing a 30-min stale batch job with sub-second stock queries Built an IaaC job management framework in Pyspark + GitHub Actions governing 200+ Databricks jobs, eliminating 100% of manual overhead and cutting onboarding from days to under 2 hours Engineered an Anomaly Detection System (Pyspark+ Z-score + Redis baseline caching + Delta Lake telemetry) that reduced pipeline incidents by 50% and MTTR from 90 → 30 minutes Delivered a Data Quality & Schema Governance Framework that saved engineers 3–4 hours per incident across gold-layer planning tables Achieved a 4.5× query speedup and 10–15% Databricks cost reduction through system-wide Spark optimisation Generated $200K in client revenue at Tredence via a Python XML parser that automated data discovery for Fortune 500 onboarding Stack I work with: Streaming & Batch: Kafka · Flink (PyFlink) · PySpark · Spark SQL · Airflow Storage & Warehousing: ClickHouse · Delta Lake · Snowflake · Redshift · S3 Cloud: AWS (primary) · Azure · Databricks Reliability: Redis · Anomaly Detection · Data Quality Frameworks · CI/CD · GitHub Actions Domains: Quick Commerce · Supply Chain Planning · Inventory Intelligence · CPG Certifications: Gen AI Fundamentals · Databricks Certified Data Engineer Associate · SnowPro Core · Azure DP-900 📩 Let's connect if you're building fast data systems or tackling ops at scale. 📊 Book a 1:1 session: topmate.io/ashish_kumar_de

Location

Bengaluru

AI stack

claude code

claude

codex

cursor

gemini

grok

Work experience

Data Engineer at Zepto, Mar 2025 - Present · 1 yr 7 mos, Bengaluru, Karnataka, India, Own Zepto's Planning data platform driving automated replenishment across FMCG, SuperStore, Cafe & Milk supply chains for 1100+ dark stores. Architected real-time Inventory Tracking System (Kafka → Flink → ClickHouse) replacing 30-min batch staleness with sub-second stock queries Built PR-based IaaC framework governing 200+ Databricks jobs — eliminated 100% manual tracking, cut onboarding from days to 2 hours Engineered Anomaly Detection System (Z-score + Redis + Delta Lake) — 50% fewer incidents, MTTR 90 → 30 min Re-architected FMCG pipelines in PySpark — 4.5× query speedup, 10–15% Databricks cost reduction, 40+ GB memory freed Built Data Quality & Schema Governance Framework — saved 3–4 hrs/incident across gold-layer planning tables Kafka · Flink · PySpark · Databricks · ClickHouse · Redis · Airflow · Delta Lake · S3

Data Engineer II at Nielsen, Sep 2024 - Feb 2025 · 6 mos, Bengaluru, Karnataka, India, Refactored and optimised ELT pipelines for Nielsen's TV audience measurement platform processing 2M+ records daily. Resolved 20% processing bottleneck via broadcast joins, salting, and multi-threading on S3-backed Spark clusters Improved end-to-end efficiency by 30% through SQL predicate pushdown and materialised Parquet intermediate layers

Data Engineering Consultant at Tredence Inc., Jan 2024 - Aug 2024 · 8 mos, Bengaluru, Karnataka, India · Hybrid, Delivered data engineering across Fortune 500 engagements — cloud migration, pipeline optimisation, and automated tooling. Built Python XML parser cutting discovery time 25%, generating $200K in new revenue Migrated IBM DataStage pipelines to ADF + ADLS Gen2 + Databricks — 20% faster processing, full Azure Monitor observability Tuned PySpark on AWS EMR across 15 TB / 12B records — replaced UDFs with Spark SQL window functions, 20% runtime improvement Designed star-schema models in Hive and Snowflake supporting BI and self-service analytics PySpark · AWS EMR · Azure Data Factory · Databricks · Snowflake · Hive · Delta Lake

Analyst - Data Engineer at Tredence Inc., Jun 2022 - Dec 2023 · 1 yr 7 mos, Bengaluru, Karnataka, India · Hybrid, Started as a campus hire, ramped quickly into production data engineering across AWS, Azure, and GCP client engagements. Built and maintained ETL pipelines ingesting from AWS S3/Glue, Azure ADLS, GCP GCS, Oracle, and SFTP across JSON, CSV, and Parquet formats Performed deep-dive SQL and Spark SQL analysis to resolve SIT/UAT defects and maintain production stability Promoted to Data Engineering Consultant within 18 months based on delivery impact Python · SQL · PySpark · Spark SQL · AWS · Azure · JIRA

Training & Placement Representative at Centre of Training & Placement,NIT Jalandhar, Sep 2020 - Jun 2022 · 1 yr 10 mos, Jalandhar, Punjab, India

Education

Bachelor of Technology - BTech, Chemical Engineering, Dr B R Ambedkar National Institute of Technology, Jalandhar

Senior Secondary | CBSE, Adarsh Vikas Vidyalaya