Staff Software Engineer · Nium

Sai
Vasanth

I'm a

Full-stack roots. Data platform depth.
I build backend systems and data infrastructure
that hold up in production, at scale, under audit.

See My Work Say Hello

About

I own problems,
end to end

I'm a Staff Software Engineer (Data Platform) at Nium, building the data platform that powers fintech operations - compliance reporting, client analytics, and production reliability. I started as a full-stack developer, grew into backend systems, and now live at the intersection of data engineering, cloud infrastructure, and observability.

I don't wait to be handed a problem. I find it, design the solution, align the stakeholders, and ship it. Whether that's a monitoring framework that catches failures before anyone escalates, or a regulatory pipeline that auditors sign off on.

99% Escalations Reduced

7+ Years in Production

Fun fact

Deleted 5 billion records from production.
On purpose. Nothing broke.

Beyond the code

I'm a stock market and geopolitics news buff - always three tabs deep in something happening in the world. I consume podcasts and music at an embarrassing rate, and I'm a bathroom singer. My favourite place is anywhere with my favourite people. I have an involuntary habit of rhyming things mid-conversation - my close friends won't let me live it down. Work should feel like your favourite video game. If it doesn't, something's off.

Download Resume

Experience

Where I've
Worked

Aug 2023 - Present Hyderabad

Staff Software Engineer

Nium

Built a custom MCP server connecting AI assistants to 20+ Redshift databases, cross-database joins, PII masking, and multi-agent validation workflows. Turned hours of manual data validation into an AI-coordinated pipeline during development phase.
The data team had no alerting - every pipeline failure landed as a Slack escalation at 2 AM. I built the entire observability stack from scratch: Datadog, Loki, PagerDuty, Grafana, Prometheus. 99.99% accuracy over 2.5 years. Escalations down 99%. The on-call now knows before anyone else does.
Regulatory reporting was running on legacy Pentaho jobs nobody fully understood. I took ownership of the Safeguarding migration for EU, UK, SG - read the legacy SQL line by line, pulled the right people into the room, and rebuilt it on Redshift. Compliance teams went from manual Excel checks to reliable dashboards. Any change now takes days, not sprints.
Built a config-driven, multi-threaded S3-to-Redshift copy framework from scratch. No code change per new use case - just config. Handles millions of records, de-duplication, and parallel ingestion. It's now the backbone of core regulatory and analytical workloads. Presented it at TGIT - got personal appreciation from the Director.
When an audit needed MasterCard transaction monitoring data backfilled from Jan 2024, the source team said 10 days. We delivered ~21M records in 2.5 hours. Zero exceptions. Audit closed successfully.
Led the org-wide GenAI evaluation - compared Copilot, Cursor, and WinSurf on live engineering use cases from our own codebase. Engaged directly with CodeAnt's CEO. Recommended GitHub Copilot based on real evidence. Secured 50 licences for pilot rollout.

PythonAWSRedshiftTerraformDatadogPagerDutyLokiDMSGlueLambda

May 2021 - Aug 2023 Hyderabad

MTS-2

Unravel Data

Took full ownership of the BigQuery pipeline - undocumented, single-threaded, built on fragile ObjectNode parsing that made testing nearly impossible. Refactored the entire codebase to type-safe POJOs, ported it into a dedicated Kafka worker with independent partitions, and scaled it to handle hundreds of concurrent GCP projects. Drove it from internal V1 to public release, shipping the observability platform across 4.7.5.0, 4.7.7.0, and 4.8.1.0. Trained the full cross-pillar team and owned every architectural decision end-to-end. Won the Above and Beyond Award for this body of work. The platform was covered by Datanami.
The night 4.8.1.0 shipped, a critical production issue hit. No handoff, no playbook. I debugged the failure, validated the fix across 10 customer clusters, and drove a clean build by morning.
GCP project onboarding required 10 minutes of manual steps. I built a hackathon prototype that automated it end-to-end. It shipped as a core product feature in 4.7.5.0 - now the entry point for every new Unravel customer.
Led the Snowflake pipeline from zero - requirements, data model, and initial delivery in 4.7.4.0. Defined from scratch how customer-environment metrics mapped to Unravel's schema.
Was the sole Kafka SME. When production failures came in, I reproduced the exact customer environment locally and debugged at Java class level to trace the root cause - keeping enterprise pipelines stable without escalation.

JavaPythonKafkaBigQuerySnowflakeElasticsearchOpenSearchGCPPostgreSQL

June 2019 - April 2021 Hyderabad

Senior Product Developer

Inncircles

Led the TechLearn edtech platform - an internal product for live online learning. Integrated the Zoom Web SDK end-to-end, letting users schedule, host, and join webinars without leaving the app.
No security framework existed - access control was scattered across products. I designed and built the company-wide RBAC permissions framework from scratch. It became the single access control foundation across all products.
Users in low-connectivity areas were getting logged out mid-session on a 6-hour token expiry. I built a refresh token framework supporting both web and mobile simultaneously, extending sessions to 30 days. Rolled out company-wide.
API responses were inconsistent across services, causing constant frontend integration bugs. I designed and implemented a unified response and error-handling middleware - standardised all services, cut error rates by 40%.
Led the in-house ticketing and support framework (similar to Jira) used by clients and internal teams. Mentored 3 engineers while delivering it end to end.

MEAN StackNode.jsAngularMongoDBExpress.js

Skills

What I Work
With

Languages

PythonJavaJavaScriptTypeScript

Cloud & Infrastructure

AWS GlueLambdaStep FunctionsDMSS3RedshiftRDSGCP BigQueryGCP Pub/SubTerraformJenkins

Databases & Warehouses

RedshiftBigQuerySnowflakePostgreSQLMongoDBMySQLElasticsearchOpenSearch

Observability

DatadogPrometheusGrafanaLokiPagerDutyCloudWatch

Backend & APIs

Node.jsExpress.jsREST APIsApache KafkaMEAN Stack

Practices

System DesignTechnical LeadershipCross-team CollaborationMentoringCode ReviewAPI Design

Selected Work

Things I've
Built

Nium 99% fewer escalations

Monitoring & Alerting Platform

Designed the entire observability stack for the data team from scratch — Datadog, Loki, PagerDuty, Grafana, and Prometheus. 99.99% alert accuracy sustained over 2.5 years. Reduced production escalations by 99%.

PythonTerraformDatadogLokiPagerDutyGrafana

Nium Millions of records at scale

Config-driven Copy Framework

Built a multi-threaded, config-driven S3-to-Redshift ingestion framework from scratch. No code changes per new use case — just config. Powers regulatory reporting and core analytical workloads. Presented at internal TGIT.

PythonAWS S3RedshiftMultithreading

Unravel Data Featured in Datanami

BigQuery Observability Platform

Redesigned the BigQuery data pipeline and model, migrated search from Elasticsearch to OpenSearch, and drove three major product releases. The platform was featured in Datanami as a key revenue driver.

JavaPythonBigQueryOpenSearchKafkaGCP

Contact

Let's Work
Together

Open to interesting engineering challenges, leadership roles, and good conversations. Whether it's a new opportunity or just a question — my inbox is open.

vasisai128@gmail.com

SaiVasanth

I own problems,end to end

Where I'veWorked