Independent cloud architect, working with startups, SaaS companies and cloud partners across India. My background is carrier-grade infrastructure — the kind where downtime is measured in minutes per year, not hours per month. I bring that discipline to teams who've outgrown whatever they set up in year one.
Two weeks. I go through your AWS, Azure or GCP bill line by line — idle instances, oversized databases, storage nobody owns, reservations you should have bought — and hand back a ranked list of changes with the monthly saving next to each. Most teams find 20–40% in the first pass.
Discovery and dependency mapping, target architecture, migration waves, cutover and rollback plan — then the execution. Including the parts most plans skip: what stays on-prem, what gets rewritten rather than lifted, and what the monthly run cost actually looks like afterwards.
Veeam or Rubrik architecture sized to an agreed RPO and RTO, with immutable copies, offsite replication and a documented recovery runbook. Ends with a real restore test, because a backup nobody has restored from is a hypothesis.
Production account structure, networking, identity, CI/CD — all in Terraform, handed over with runbooks. For teams who've outgrown the console and the one engineer who knows where everything lives.
Cluster design, GitOps delivery, autoscaling, observability and the security defaults nobody sets at the start. Greenfield build, or the rescue of a cluster that became load-bearing by accident.
A fixed block of hours each month for design reviews, vendor decisions, post-mortems and the calls your team shouldn't be guessing on. For companies who need senior architecture judgement without a senior headcount.
I take architecture workstreams on subcontract for cloud partners and managed service providers — SOW-based, white-labelled where needed, working inside your delivery process and reporting to your engagement lead.
Telco and OpenStack-based private cloud delivery, where uptime targets aren't aspirational and capacity planning isn't a guess. That's the standard I bring to much smaller systems.
AWS, Azure and GCP — multi-account design, networking, identity, migration planning, and an honest answer when someone asks which one to pick.
Kubernetes, Helm, GitOps pipelines and Terraform — the automation that makes an environment reproducible instead of remembered.
Veeam and Rubrik — backup architecture, immutable and air-gapped copies, replication, and DR designed to a stated RPO and RTO rather than an assumed one.
On-prem to cloud: discovery and dependency mapping, target architecture, wave planning, cutover and rollback, and a run-cost model you see before you commit.
FinOps practice, capacity planning, SLOs, and monitoring that tells you something is wrong before a customer does.
Reference implementations, built and measured end to end to demonstrate approach. Architecture and decisions walked through on request.
Account and subscription structure, hub-and-spoke networking with a documented CIDR plan, SSO-based identity with no long-lived keys, and policy guardrails — all in Terraform. The headline number is how long a new workload environment takes to provision once the pipeline exists, replacing a ticket queue that usually runs into weeks.
Managed cluster with GitOps delivery, autoscaling at both levels, full observability and deny-by-default network policy. I killed a node mid-deployment, measured recovery, fixed what the test exposed, and ran it again — the headline number is recovery time after that fix.
A deliberately wasteful estate — oversized instances, always-on dev, orphaned volumes, on-demand pricing for steady-state load — audited the way I'd audit a client's. Output is a ranked table of changes with monthly saving, effort and risk against each.
A cloud architect working across public and sovereign cloud environments, leading transformation, migration and solution architecture work end to end — workload assessment and dependency analysis, HLD and LLD, implementation, and production cutover.
Much of that has been public-to-sovereign migration for regulated environments, where the architecture has to satisfy security, availability and performance requirements that are written down and audited rather than assumed. Alongside the build work sits the commercial side: customer discovery, technical workshops, RFP responses, BOQ and BOM, effort estimation and sizing. It means the proposal you get from me reflects what the work will actually take.
Particular interests: zero-downtime architectural change on live production systems, autoscaling designs built on golden images and controlled instance refresh, cost forecasting and capacity models that inform commercial decisions, and backup and DR designed to a stated RPO and RTO. I also write my own tooling where it saves client time — inventory and cost extraction that turns a multi-day assessment into about an hour.
Tell me what you're running and what's going wrong with it. I reply within a day.
Add your scheduling link (Cal.com or similar) and your LinkedIn here.