Skip to content
Nitin Saini

Data operations for products built on external data.

I get new data sources live fast and keep them trustworthy as they change: freshness, coverage and accuracy, measured.

See the workGet in touch

Many sources in. One trusted dataset out.

Quality gate

remapped · released

Trusted dataset

customers · AI models

event log · select a feed to inspect its mapping

  1. 14:09Batch released; schema drift 0
Mapping the feeds is the easy part. Keeping them trustworthy when they change is the job. Illustrative data.

Animation: three data feeds with different field names are mapped into one shipment dataset while quality checks run. One feed then renames a field without notice. The schema change is detected, the affected batch is held at a quality gate so published data stays correct, the mapping is updated and the batch is released. No customers are affected. The data is illustrative.

Experience at

  • Compass
  • USAA
  • Oracle Financial Services
  • TrackSo (co-founder)

How I think

  1. “The data looks fine” isn’t a metric.

    Freshness, coverage, drift and accuracy can all be measured. Until they are, data quality is an opinion.

  2. Late beats wrong.

    When a feed breaks, hold the data and say so. A delay costs minutes. Wrong data costs a customer’s trust.

  3. AI-ready data is well-run data.

    There’s no shortcut. A model inherits every gap, delay and silent schema change in the feeds underneath it.

What it looks like in practice

All the work

Compass · Senior data analyst

50%+

less customer impact from data issues

Monitored the pipelines behind a customer-facing platform, measured freshness and availability, and fixed problems before customers saw them.

Read the story

USAA · Senior analyst

80%

less time validating data

Automated data-quality validation with Python and SQL, and added quality checks to the pipelines behind four reports.

Read the story

TrackSo · Co-founder

10x

onboarding capacity, to 5,000+ assets

Energy-asset data platform processing about 500K live records a day. I built the roadmap and the analytics engine; recurring customers rose 70%.

Read the story
Portrait of Nitin Saini

About

I trained as an electrical engineer, spent more than two years checking data quality for more than 25 large banks at Oracle, then co-founded an energy-data startup. Since then I’ve led data quality and operations work at USAA, Compass and now REalchemy. The thread through all of it: messy data from many sources, made dependable.

Career

  1. 2022–nowREalchemySenior Director, Data
  2. 2021–22CompassSenior data analyst
  3. 2020–21USAASenior analyst
  4. 2016–19TrackSoCo-founder
  5. 2013–15Oracle Financial ServicesTechnical analyst

Education

  • Master’s in Data AnalyticsMcDaniel College, 2025
  • M.S. in Business Analytics and Project ManagementUniversity of Connecticut, 2020
  • B.E. in Electrical EngineeringDelhi College of Engineering, 2013

Topics I can talk about

  • Onboarding third-party data at speed
  • Measuring data quality: freshness, coverage, drift, accuracy
  • The data layer under AI: what “AI-ready” really takes
  • From founder to operator: building data teams in ambiguity

Press kit

Bios and a headshot for event organizers, podcasts and expert networks.

Working on a feed problem?

Tell me about the feeds and what’s breaking. It’s a conversation, not a pitch.

Get in touch