Vikram Babariya

Designed and delivered a production-grade zero-trust AWS serverless platform (live at https://vikram-sre.dev/) — 100% declarative Terraform IaC across five modules, zero static credentials, shift-left security gates at every layer.

• Eliminated long-lived IAM credentials across 100% of deployments via GitHub Actions OIDC federation — short-lived STS tokens scoped to least-privilege, with cross-repo and cross-branch token reuse mathematically prevented by trust policy sub claim binding.
• Replaced all ClickOps with a declarative Terraform codebase across five focused modules (state-backend, cdn, compute, dns, iam) — every AWS and Cloudflare resource version-controlled and peer-reviewed, DR recovery reduced to a single terraform apply.
• Architected a five-gate Terraform CI/CD pipeline (fmt → validate → checkov SAST → plan → apply) — HIGH/CRITICAL findings block merge, intentional waivers documented inline per ADR, and terraform plan output posted as an idempotent PR comment for peer review before any resource change.
• Enforced Shift-Left quality gates across the full pipeline — ajv-cli/yamllint "Fail Fast" contracts block invalid artifacts pre-build; post-apply smoke tests verify seven live infrastructure invariants including IAM Policy Simulator assertions confirming PoLP boundaries hold.
• Hard-capped the full AWS stack at ₹500 INR/month via dual-threshold budget alerts (35% actual / 100% forecasted) wired to SNS — API Gateway rate limiting guards against Layer 7 Denial-of-Wallet attacks.

• Architected a Django-based Unit Resolution module for an enterprise Register Map Platform, automating analog-to-digital data translations while integrating strict API quality gates and comprehensive server-side logging to guarantee data integrity.
• Engineered a Node.js orchestration engine to dynamically parse and execute Python test vectors within managed, ephemeral virtual environments, implementing runtime call-stack inspection to enforce "Fail Fast" validation on I2C/UART hardware operations.

• Architected a scalable Wafer Visualization Platform by integrating Snowflake with Django, implementing multi-threaded data ingestion to process 100+ massive datasets concurrently and reducing defect analysis time from hours to minutes.