Cloud delivery on AWS
Setting up production on AWS in a shape that repeats: ECS and ECR, infrastructure as code, one CI pipeline shape, and a security baseline (encryption with KMS, WAF, private networking, Secrets Manager) in place before sensitive data arrives. That includes migrations, such as moving a 30 GB production MySQL database to RDS with DMS and no downtime.
DevOps and SRE practice
Building the practice from one engineer and one pipeline. Releases went from once a week to 3 to 6 a day in dev and 1 to 2 a day in production, and P1 incidents from several a month to near zero by 2022. Quality gates with Jest, Cypress and SonarCloud on every active project.
Document AI and RAG infrastructure
RAG-based search over legal documents, a Whisper transcription pipeline for video evidence, and docling-batch-extract: batch PDF to Markdown on one CPU-only server, sized from measured benchmarks and compared on cost with cloud OCR APIs.
Developer tooling and AI-assisted engineering
CI/CD with GitHub Actions and AWS CodePipeline, and building software with Claude Code and GitHub Copilot the way I'd run any project: versioned plans with requirement IDs, an acceptance check for every requirement, a go-ahead gate before release, and prompts that get better after every successful run. The practices are written down in public.
Engineering leadership
Founding engineer at a company that grew from 14 to 150+ engineers. Teams of 4 to 8 engineers across four concurrent client engagements, with architecture, delivery and the client relationship owned end to end, and no separate project management layer.
Mentoring
Industry mentor with the IEEE India Council (Student Mentorship Program 2024, Career Mentoring Program 2025): mock interviews, architecture walkthroughs and career clarity sessions for final-year engineers. One-to-one sessions on Topmate for career guidance and for IEEE mentees.