Pavan Kumar
Singara.
I build backend systems, data pipelines & full-stack apps that scale.
Software Engineer II at the State of New Mexico, with 2+ years of experience shipping reliable, well-tested systems across Python, Java, and JavaScript. I own features end-to-end and thrive at the intersection of architecture and execution.
processing
per day
friction cut
served
About Me
Hey! I'm Pavan, a software engineer based in Albuquerque, NM who loves building things that actually work — reliably, at scale, and with clean code underneath. I graduated from the University of New Mexico with a B.S. in Computer Science and a minor in Business & Entrepreneurship.
My work spans the full stack: designing Java REST APIs and Python batch pipelines for statewide government platforms, implementing Kafka-based event systems for high-volume API observability, and securing enterprise infrastructure under HIPAA and ISO 27001. I'm as comfortable in a CI/CD pipeline as I am in a Splunk dashboard or a React SPA.
Outside engineering, I've published ML research applying kernel mean embedding to predict elastic constants of multi-component alloys — bridging materials science and machine learning.
I'm currently open to new opportunities. If you're building something interesting, let's talk.
Where I've Worked
- Built a full-stack statewide benefits platform — Java REST APIs handling eligibility logic and audit trails, Python batch pipelines for data validation, and a Razor Pages frontend serving warehouse staff and executives.
- Automated data reconciliation across multiple downstream state systems, reducing manual processing effort by 30% and improving cross-system accuracy.
- Designed and optimized SQL Server stored procedures, views, and indexes that significantly improved reporting performance for high-traffic queries.
- Built and managed Airflow DAGs to orchestrate automated data workflows and batch processing jobs across multiple state systems.
- Maintained CI/CD pipelines with full rollback support on a legacy .NET 4.6 codebase; coordinated multi-environment deployments across teams.
- Rebuilt the Developer Console using OAuth & Spring Boot microservices, adding observability across 2M+ daily API calls and reducing sign-up friction by 80% through Okta and Auth0 federated SSO.
- Implemented Kafka-based event pipelines for high-volume message routing; built an audit microservice tracking user activity across 5 geographic regions.
- Integrated AngularJS frontend with Node.js backend, improving performance by 35%; managed Kubernetes deployments and centralized logs via ELK stack for ~700,000 users.
- Built and maintained Splunk dashboards and alert pipelines monitoring security events across state infrastructure; configured Palo Alto firewall rules and conducted real-time threat log analysis.
- Ran enterprise vulnerability assessments under ISO 27001 and HIPAA, correlating Nessus scan results with SIEM data to validate patching and safeguard PHI/PII.
- Standardized Splunk Universal Forwarder deployment across Linux and Windows environments, improving SIEM coverage and incident detection speed.
- Implemented Auth0 SSO reducing access management overhead by 30%; built ReactJS SPAs and refactored legacy JavaScript to TypeScript for improved maintainability.
- Architected microservices within an Istio service mesh on Kubernetes; automated AWS infrastructure via CloudFormation for EC2, Lambda, and API Gateway.
- Integrated Airflow with AWS S3 and Google Cloud Storage for scalable data ingestion pipelines; wrote Python scripts cutting debugging time by 40%.
Skills & Stack
Things I've Built
Applied kernel mean embedding to predict elastic constants of multi-component alloys. End-to-end ML pipeline: data collection, feature engineering, model training, and rigorous evaluation. Bridges computational materials science and applied ML.
Real-time auction bidding application built with Java and JavaFX. Features live bidding mechanics with concurrency control, item catalog management, bid validation, and a responsive desktop GUI. Handles simultaneous bidders with conflict resolution.
Location-aware full-stack web application helping UNM students find campus food and basic needs resources. Built with geolocation APIs, an interactive resource map, and an accessible mobile-first interface designed to reduce food insecurity on campus.
Low-level systems tool that injects controlled memory errors to validate hardware error-correction (ECC) mechanisms. Used for stress-testing memory reliability under fault injection conditions — useful in high-availability and embedded systems contexts.
What's Next?
Let's build something great.
I'm currently open to new opportunities — whether it's a full-time role, a contract, or just an interesting conversation about systems and scale. My inbox is always open.
Say Hello ↗