Skip to content
Shivam Datta
Talk to Shivam

About

Professionally suspicious of good-looking numbers since 2014.

Who

Shivam Datta

Shivam Datta · Luxembourg

I started in 2014 as a decision scientist at Mu Sigma, where clients arrived with fuzzy questions and left with models. I moved into performance marketing at MiQ, spent three years consulting at Eli Lilly, and then joined Amazon in Luxembourg to lead analytics for its safety, compliance and risk teams.

In 2026 Amazon closed my whole team in a restructuring. I took it as a nudge to do this for myself.

I work AI-first, so things get built fast. What you’re really paying for is judgement: what to build, how to prove it works, and what to quietly throw away.

Talk to Shivam

Where I’ve worked

Select a role to see more

2021 – 2026
LuxembourgAmazonBusiness Analyst ManagerAnalytics and reporting for Global Safety & Risk Control: 500+ training and compliance programmes across North America, Europe and India.
  • Built one dashboard for dangerous-goods, food-safety and trade-compliance training: 87,000+ evaluation responses and 16,000+ submission scores, in Amazon QuickSight.
  • Rewrote how scores are calculated (averages instead of maximums, four-decimal precision) and rebuilt the training lookup as learning path, course, training. Its build time dropped by 40%.
  • Found 83% of trainee records missing from a weekly report’s source, then led the redesign through to go-live: 358 records became 2,158+, with 200+ columns mapped and 72 core fields defined.
  • Scoped the fix for an audit finding covering 3,600+ employees: five layers of validation with daily cross-system checks, and aviation-standard validity rules (24-month recertification, 90-day expiry alerts).
  • Pulled data from 10+ systems into one dataset, with 73+ column mappings and de-duplication, and cut the core query from 400+ lines to about 100.
  • Automated the recurring reporting: 260+ hours a year of manual work gone.
  • Put an AI chat agent on the dashboard for plain-language questions and for reading the mood of training feedback, and led the recovery of 300+ deleted training packages, with an AI knowledge assistant over the rescued content.
  • Stack: SQL (Redshift, Athena, PostgreSQL), Amazon QuickSight, Python, AWS.
2018 – 2021
BengaluruEli Lilly and CompanySenior Associate ConsultantAnalytics for the US business: how products were performing across channels, for the brand and sales teams.
  • Owned the analytics for a product launch in an emerging market: who to target, and whether it was working.
  • Automated product and payer performance tracking from end to end, saving each analyst 30+ hours a month.
  • Built an Automated Insights Generator on my own: give it a dataset and it returns the charts and the findings in minutes.
  • Ran data diagnostics and visualisations that the business intelligence teams used for product planning.
  • Stack: SAS, SQL, Python, R, Tableau, advanced Excel and VBA.
2017 – 2018
BengaluruMiQ DigitalSenior AnalystPerformance marketing in programmatic advertising, for advertisers and their agencies.
  • Wrote insight reports before, during and after each campaign: trends by time, geography, device, ad placement, domain, keyword and audience segment.
  • Optimised a wide range of live display campaigns for delivery and margin.
  • Built conversion-likelihood models for UK businesses’ online campaigns, checked against the people actually targeted, to estimate the extra revenue.
  • Processed campaign data on AWS: custom clusters, S3 and Redshift.
  • Built the pitches that won and won back clients, from keyword research and a hard look at their online presence.
  • Stack: AppNexus, DoubleClick Bid Manager, SQL (Redshift, Hive), AWS, SAS.
2014 – 2017
BengaluruMu SigmaDecision ScientistDecision science for a pharma manufacturer and an airline, with three teams to lead.
  • Built an early-warning system for a Fortune 50 pharma client: a time-series model (ARIMAX) forecasting four to eight weeks of each physician’s performance, to flag the ones drifting off course.
  • Introduced a gradient-boosting model to predict future losses from claim duration, submission delay and case ratios, replacing the traditional method.
  • For an airline: measured campaigns with test-and-control matching, and modelled who was most likely to book so the marketing could be aimed at them.
  • Built launch trackers for new products, analysed why patients and physicians were dropping off, and studied how one class of drugs eats into another’s market share.
  • Was solely responsible for the weekly and monthly brand reports and the master datasets behind them, rebuilt as reusable, automated code.
  • Led a ten-person BI team for the pharma client and two analytics teams for the airline. Pitched new work to clients, worth over US $1M in savings and avoided risk.
  • Stack: SAS, SQL, R, Excel and VBA, Spotfire.

Credentials

  • B.Tech, Software Engineering — SRM University (2010–2014)
  • AWS Certified Solutions Architect
  • DeepLearning.AI Machine Learning Specialization