Harshith
Software Engineering Intern
Talent
HyderabadMember since August 25, 2026
SQLSpring BootNode.jsReact.jsNext.jsTailwind CSSDockerKubernetesAWSEKSIAMCloudFrontSecrets ManagerPulumiGitCI/CDGrafanaMySQLMongoDBJWTRBACGoogle OR-ToolsCP-SATSoftware EngineeringOOPsGitOps
·
Hire this person
Send a job offer directly to this candidate
SQLSpring BootNode.jsReact.jsNext.js
About
Automated secret rotation via ESO + AWS Secrets Manager, eliminating manual credential overhead across 5+ multitenant Kubernetes environments. Provisioned isolated testing infrastructure using Pulumi IaC on AWS, reducing setup time by 50% and enforcing continuous compliance loops. Maintained cloud-native CI/CD and GitOps pipelines for automated deployments, increasing release cadence by 20% while mitigating deployment failures.
Experience
Software Engineering Intern at HugoHub (2026-01 – 2026-06)
- Automated secret rotation via ESO + AWS Secrets Manager, eliminating manual credential overhead across 5+ multitenant Kubernetes environments.
- Provisioned isolated testing infrastructure using Pulumi IaC on AWS, reducing setup time by 50% and enforcing continuous compliance loops.
- Maintained cloud-native CI/CD and GitOps pipelines for automated deployments, increasing release cadence by 20% while mitigating deployment failures.
- Hardened AWS IAM security postures across EKS, Cognito, and CloudFront using Least Privilege Access, while constructing 10+ Grafana dashboards for real-time telemetry.
Software Engineering Intern at CrossBo (2026-06 – 2026-06)
- Architected a production-scale housekeeping plan optimizer on Google OR-Tools CP-SAT: replaced an O(M · N²) route model (28s → 0.7s) with a decompose-and-validate solver planning 50 attendants × 300 units in ∼10s.
- Encoded duty-of-care safety invariants (lone-worker prevention, co-presence, zone controls) as hard CP-SAT constraints, preventing 100% of structural safety violations upon scheduling.
- Hardened execution pipelines through up-front input validation covering 4 error categories (inverted windows, cycles, locks), returning deterministic coded errors.
- Cut overall solve latency by 75% by reusing solved models for infeasibility diagnosis (∼40s → single Solve()) and indexing task matrices to achieve linear time complexity (O(tasks²) → O(tasks)).
Education
B.Tech in Computer Science – Indian Institute Of Information Technology, Lucknow (IIITL) (2026-06)
Skills
Software Engineering Intern



