Engineering Lead · Ollion
Amit Sharma — Cloud & Platform Engineering Leader
Zero-downtime cloud migrations, SRE-grade Kubernetes platforms, and the engineering programmes that ship them.
I design resilient cloud platforms and lead the engineering programs that ship them. From SRE-grade Kubernetes to zero-downtime migrations, I own delivery end-to-end across GCP, AWS, and Azure.
What I Stand For
Principles that shape how I work
Zero downtime is a constraint, not a goal
Teams that achieve zero downtime design for it from day one. It's not something you bolt on at the end — it shapes every architectural decision, from dual-run strategies to rollback triggers.
Migrations are org problems first
The technical challenges are rarely what derails a migration program. Misaligned stakeholders, unclear ownership, and InfoSec teams looped in too late — these are the real risks.
Platform engineering is about enabling autonomy
A good platform team makes itself invisible by enabling other teams to move fast safely. The goal isn't to run Kubernetes — it's to make the 100 teams using it not need you.
Selected Work
Case studies with real outcomes
QuotaLens — Local Claude Quota Monitor
Open-source, local, self-hosted monitor for Claude Pro/Max subscription quota. Polls every minute into a SQLite file you own and forecasts whether the current session window will exhaust before it resets. Dashboard, Prometheus metrics, burn-rate webhooks. v1.0 on PyPI.
v1.0 released on PyPI; MIT open-source; macOS, Linux and Windows tested in CI
Joymunity — Employee Engagement Platform
In development. Joymunity helps organisations build joyful workplaces through AI-powered employee-engagement experiences. Next.js 15 web app, FastAPI backend, Terraform-managed infrastructure.
In development — architecture and core platform underway
Kubernetes Platform Modernization for a Securities Trading Platform
Led platform modernization, Kubernetes migration, and security hardening for a securities brokerage's mission-critical trading and internal platforms across GCP and Nutanix. Migration and security hardening complete; continuing as advisory.
~100 applications onboarded to Kubernetes with zero unplanned downtime
Mobile Ad Platform: Azure to GCP Migration at 4.5M Req/Sec
Led end-to-end migration of a mobile advertising platform's high-throughput customer workloads from Azure to GCP — zero unplanned downtime, 4.5M+ requests per second, full program governance.
4.5M req/sec platform migrated with zero unplanned downtime
Koo: AWS to GCP in 4 Months, Zero Downtime
Migrated Koo's entire production platform — 40+ app services, 48 ML services, and 21 databases — from AWS to GCP within a strict four-month timeline with zero downtime.
40+ services, 48 ML services, 21 databases migrated in 4 months with zero downtime
Tokopedia: AliCloud to GCP — Migration Hero
Led zero-downtime migration of Tokopedia's AliCloud workloads to GCP, earning the 'Migration Hero' recognition for execution excellence across one of Southeast Asia's largest e-commerce platforms.
Zero-downtime migration recognized with 'Migration Hero' award
Latest Thinking
Writing from two angles
Insights
Technical & industry perspective
Ethan Hunt Would Blow Your Error Budget
Movies need a protagonist who goes dark, works alone, and saves the world with one second left on the clock. Incident management needs the exact opposite. Here's what the silence between first responders and stakeholders actually costs.
The Best Retention Metric Is Who Follows You and Who Outgrows You
Conventional retention metrics measure who stayed. The more honest measure is who joined your next program voluntarily, and who left because they grew into something bigger than you could offer. Both are good outcomes.
`count` Is a Production Outage Waiting for a Deletion
Terraform's `count` assigns resources by index. Delete the middle item and the indices shift — Terraform destroys and recreates everything after it. On subnets, node pools, or persistent volumes, that's not a warning. It's an incident.
Default-Deny Starts in Observe Mode
Kubernetes NetworkPolicy default-deny is the right posture — but enforcing it cold in production breaks applications and creates an incident. The right approach: observe first, map actual traffic, enforce progressively. Same staged approach as a migration.
Disk Full at 10am: Logs Are an Architecture Decision
Log storage is treated as an afterthought — set a retention, set a volume, deploy. It becomes an architecture decision when the disk fills and the logging agent takes down something else. Logs need their own failure modes designed in.
Estimates Are Architecture: Writing SOWs That Leave No Loose Threads
A Statement of Work is architecture for delivery. The same failure modes apply: if a dependency is undocumented in the SOW, it will bite you at the worst moment. The discipline is to write the SOW as if the project has already gone wrong.
Perspectives
AI, society & the human dimension
The Independence Gap: Why Control Masquerades as Care in Indian Households
Patriarchal oversight and delayed autonomy are quietly undermining independence and decision-making skills in young adults.
Why Ancient Rituals Are Better Productivity Tools Than Your Todo App
How traditional structures from Indian culture offer surprising frameworks for modern focus and goal attainment.
When Engineers Stop Talking – The Cost of Silence
How speed, remote work, and AI have eroded conversation—and what we're losing in the process.
The Competitive Trap: Why India's Education System Is Quietly Killing Innovation
How competitive obsession in Indian education undermines technical excellence and collaborative problem-solving.
The Engineer Who Isn't Afraid
AI can now write code, generate infrastructure, and answer architecture questions. Most of the engineers I know are scared. I think the scared ones are asking the wrong question.
From the Lab
Self-hosting and systems experiments
Replacing Google Photos with Immich
After years of Google Photos, I migrated my entire photo library to a self-hosted Immich instance. The migration took a weekend. The result is faster, more private, and — surprisingly — more capable for browsing.
Private AI at Home with Ollama + Open WebUI
Running large language models locally means your prompts never leave your machine. Here's how I set up Ollama and Open WebUI on my home lab, what models I use daily, and where the limits are.
Why I Built a Home Lab on a Gaming Laptop
Running Proxmox on a gaming laptop turns out to be a surprisingly capable home lab setup — enough CPU, ECC-free RAM that doesn't matter for learning, and a screen attached. Here's why I did it and what I learned.
More on how I got here in about, or see the full career history.
Certifications
AWS & Google Cloud certified
Solutions Architect – Professional
Amazon Web Services
Professional Cloud Architect
Google Cloud
Professional Cloud Security Engineer
Google Cloud
Generative AI Leader
Google Cloud