Amit Kumar Sahu
Lead Data Engineer — building lakehouse & semantic platforms at scale
Senior / Lead / Staff roles · 9+ years · Pharma & SaaS
Senior Data Engineer with 9+ years building enterprise-grade data lakes, lakehouse platforms and semantic analytics layers across pharma (Eli Lilly, 200K+ clinical assets) and SaaS (Kenko AI, 500+ multi-tenant clients). Consistent 40–60% improvements in pipeline efficiency, data quality and cost.
45%
Query latency reduction (Kenko AI lakehouse)
500+
Multi-tenant fitness-studio clients served
200K+
Clinical assets cataloged (Eli Lilly)
98%
SLA achieved, up from 70% (Kenko AI)
How I lead
Leadership isn't a title change — it's already the job
Three examples of owning outcomes beyond my own code.
Team leadership
Led a 5-member team — AspireNXT
Owned delivery of AWS data lake pipelines across finance, industrial IoT, and healthcare clients — delivery velocity +30% — plus weekly knowledge-sharing sessions that lifted team productivity +20% and cut onboarding time −50%.
Cross-team standards
Standardised platforms for 5+ teams — Eli Lilly
Built 20+ reusable Data Products and DWS APIs for the Enterprise Data Backbone, adopted across 5+ teams — cut incident response time −40% via ServiceNow integration.
Platform architecture
Architected a multi-tenant lakehouse — Kenko AI
Designed the lakehouse and Cube Cloud semantic layer serving 500+ clients end-to-end — query latency −45%, SLA 70%→98%, BI licensing costs −35%.
Experience
9+ years across pharma, SaaS and IoT
Reverse-chronological, including a planned career break spent upskilling.
Kenko AI
Feb 2026 – Apr 2026Senior Data Engineer
SaaS Fitness Tech · Bengaluru
- Architected AWS Data Lakehouse for 500+ fitness-studio clients via DMS CDC → S3 Bronze → Apache Iceberg Silver (S3 Tables) → Athena Gold: query latency −45%, SLA 70%→98%.
- Designed Cube Cloud LDM: 18 cubes, 11 pre-aggregated views, 58 KPIs across Finance, Memberships, Classes, and Marketing; JWT-based multi-tenant RLS for 500+ operators.
- Built 9-check DQ framework with SNS alerting: incidents −60%, confidence 82%→97%.
- Delivered 20-table Gold analytics layer replacing GoodData: BI licensing −35%.
- Role eliminated in a company restructuring.
Planned career break
Jun 2024 – Jan 2026Upskilling & independent projects
Independent · Bengaluru
- Upskilling in Iceberg/Delta lakehouse architecture and GenAI on AWS Bedrock.
- Built Astra Data Platform and Amavya.
Eli Lilly and Company
Feb 2021 – May 2024Engineer II, Software Configuration and Development
Pharma · Bengaluru
- Standardised 20+ reusable Data Products/DWS APIs for the Enterprise Data Backbone (PySpark, Airflow) across 5+ teams: incident response −40% via ServiceNow integration.
- Migrated 5+ TB CTMS clinical data IMPACT→Veeva Vault: onboarding +60% across 6+ trials.
- Axon Data Catalog for 200K+ clinical assets: discoverability +50%, validation +30%.
- Won Lilly Global Ideas & Innovation Award for LillyTV (scaled to 1,500+ employees).
AspireNXT Pvt. Ltd.
Aug 2019 – Jun 2020Data Engineer
Consulting · Bengaluru
- Led 5-member team, AWS data lake pipelines (finance/IoT/healthcare), velocity +30%.
- Migrated 10+TB, latency −40%, infra cost −25%.
- Established weekly knowledge-sharing sessions: team productivity +20%, onboarding time −50%.
SpanIdea Systems
Jan 2019 – Jul 2019Senior Software Engineer
IoT · Bengaluru
- Span Park smart parking (Python, C, Raspberry Pi, PostgreSQL, Azure IoT Hub): real-time availability tracking cut parking search time −50%.
Infosys Limited (Client: Daimler AG)
Aug 2015 – Dec 2018Senior Systems Engineer
Automotive Manufacturing · Bengaluru
- Automated critical ETL pipelines with RCA and on-call support across 12+ projects: downtime −20%, automation efficiency +15%, manual effort −30%.
Education
VSSUT Burla, Odisha
B.Tech, Computer Science and Engineering
Founded ENIGMA coding club · ACM-ICPC World Semifinalist 2013
Jawahar Navodaya Vidyalaya, Sarang, Dhenkanal, Odisha
AISSCE
Case Studies
Selected work, in depth
Problem, architecture, my role and outcome for four representative engagements.
Problem
500+ fitness-studio clients needed fast, reliable analytics on top of operational data scattered across source systems, with strict per-tenant data isolation.
Architecture
AWS Data Lakehouse: DMS CDC → S3 Bronze → Apache Iceberg Silver (S3 Tables) → Athena Gold. Semantic layer on Cube Cloud with an 18-cube LDM, 11 pre-aggregated views and 58 KPIs across Finance, Memberships, Classes, and Marketing, secured with JWT-based multi-tenant row-level security for 500+ operators.
My role
Architected the lakehouse pipeline end-to-end and designed the Cube Cloud logical data model and multi-tenant security scheme.
Outcome
Query latency down 45%; SLA improved from 70% to 98%.
Problem
Clinical trial management data lived in a legacy system (IMPACT) that slowed trial onboarding, and 200K+ clinical assets were difficult to discover and validate.
Architecture
Migration pipeline moving 5+ TB of CTMS clinical data from IMPACT to Veeva Vault across 6+ trials, paired with cataloging 200K+ clinical assets in Axon Data Catalog. Also standardised 20+ reusable Data Products/DWS APIs (PySpark, Airflow) shared across 5+ teams.
My role
Engineer II, Software Configuration and Development — led the migration workstream and the Axon cataloging effort, and built the standardised Data Products/APIs.
Outcome
Trial onboarding up 60% across 6+ trials; asset discoverability up 50% and validation up 30%; incident response down 40% from standardised APIs.
Problem
Multi-tenant Gold analytics layer needed measurable, trustworthy data quality to replace an existing BI stack (GoodData) with confidence.
Architecture
A 9-check data quality framework with SNS alerting layered on top of the lakehouse, feeding a 20-table Gold analytics layer.
My role
Designed and built the DQ framework and the Gold analytics layer that replaced GoodData.
Outcome
Data quality incidents down 60%; confidence in data up from 82% to 97%; BI licensing costs down 35% by replacing GoodData with a 20-table Gold layer.
Problem
Wanted hands-on depth in modern lakehouse table formats and GenAI infrastructure beyond day-job scope.
Architecture
Independent lakehouse platform built during a planned career break (Jun 2024 – Jan 2026), applying Iceberg/Delta architecture patterns and GenAI on AWS Bedrock.
My role
Sole builder — design, implementation and deployment.
Outcome
Open-source project published at github.com/nikuamit/astra-data-platform.
Skills
Toolbox
Languages & DB
AWS
Azure
Lakehouse
Semantic / BI
Tools
Ways of working
Projects
Independent work
Contact
Let's talk
Open to Senior / Lead / Staff Data Engineer roles — Bengaluru, remote or relocation · available immediately.
Elsewhere
LinkedIn — Amit Kumar SahuBengaluru, India