Igal Schneider, cloud architect at IT Delivered in Toronto — Igal Schneider
IT Delivered, Toronto

I run and fix AWS for media, streaming and tech companies.

IT Delivered is my Toronto consultancy. I've run production AWS since 2012, starting with moving ES3 off Rackspace, and the AWS infrastructure behind Zone TV's free streaming TV service since 2016. I migrated 273 million S3 files without an outage and found a $7,000-a-month leak in an S3 bill by reading the access logs. If your AWS bill or your uptime worries you, email me.

Email info@itdelivered.ca +1 (647) 495-5275

Ways to work with me

Pick whichever fits how much help you need. Every engagement is hands-on: I do the work myself.

Monthly retainer

I run your AWS environment month to month: deployments, upgrades, cost and security reviews, certificates and access, and the fixes when something breaks.

Fits companies that need a senior cloud engineer but not a full-time hire.

Fixed-scope project

One defined job with an agreed scope: a migration, an EKS upgrade, a legacy service rebuilt, a website moved off expensive hosting, or a cost cleanup.

Fits a specific problem with a clear finish line.

Hourly, when you call

Troubleshooting, a second opinion on an architecture, or getting a stuck deploy moving again. Billed by the hour, only when you need it.

Fits teams that mostly manage on their own.

Case files

Problems from clients' live systems, and what fixed them.

$84K+saved per year

Media and streaming

An S3 bill that grew by $7,000 a month

I loaded the S3 access logs into Athena to see which objects were actually being read, then re-architected the storage tiering around that. The bill dropped by more than $84,000 a year, without touching the product.
273Mfiles moved with no outage

Media and streaming

A 273-million-file S3 migration

I ran the move with S3 Batch Operations, alongside cross-account transfers of 12 to 22 TB of video at a time on high-bandwidth instances. The service stayed up the whole way through.
80%less cluster management overhead

Streaming TV

From kops to EKS, and every upgrade since

I moved two production Kubernetes clusters off self-managed kops onto Amazon EKS, which cut cluster management overhead by 80% and costs by more than 20%. I've run every control-plane and node upgrade since then with rolling deploys and PodDisruptionBudgets, so the platform stays up while the cluster changes underneath it.

Also

A legacy analytics service, rebuilt. For a streaming TV service, I rewrote the analytics collection service on current Java, Scala and AWS SDK versions, ran it next to the old one, and switched production over with a DNS change. The old service stayed up as the rollback for a couple of days, then I decommissioned it.
An investment company's website, moved off vendor hosting. The site ran on a vendor's WordPress and frontend hosting. I rebuilt it as a static site on S3 and CloudFront, with a small Lambda for live market data, and retired the WordPress install.
An e-commerce platform moved from Google Cloud to AWS. I migrated a food-industry e-commerce platform without interrupting its customers, split its monolith into services, replaced Cloudinary with an in-house image pipeline on Lambda, S3 and CloudFront that cut image delivery costs by more than 20x, and fixed the slow MySQL queries holding up order processing.
AWS accounts and web servers consolidated. Merging several AWS accounts and web servers into fewer cut AWS costs by up to 60%.
Set-top boxes stopped playing after a TLS change. New CloudFront TLS requirements broke a fleet of older Windows CE devices. I built an HTTP-to-HTTPS proxy on Kubernetes that brought playback back in production while the long-term fix was built.
A Redshift cluster stuck at 100% CPU. I found the runaway query, then rebuilt the ad and viewing reports as pre-aggregated daily tables with proper distribution and sort keys and materialized views.
A monolith split into services. I broke one application into separate config, programming and UI services, so each team could deploy on its own schedule.

What I cover

Three areas that most companies hire three different people for.

Architecture and platform

AWS design and migrations, including moves that happen under live traffic. Production EKS: upgrades, node migrations and zero-downtime rollouts. Streaming delivery on S3, CloudFront, Cloudflare and Akamai, down to HLS playback.

DevOps and security

CI/CD on Jenkins, Bitbucket Pipelines and Bamboo, with infrastructure as code and smaller images. IAM, TLS certificate renewals, WAF rules and security monitoring, all kept in code and reviewed like any other change.

FinOps

Cost analysis from real usage data: access logs, Athena, Cost Explorer. Storage tiering and right-sizing that lower the bill without changing what your customers see.

For every client I manage EC2, S3, CloudFront, Redshift, RDS (MySQL, PostgreSQL and SQL Server), Cognito, ACM and IAM.

Who I work with

Media, streaming, health-tech, e-commerce and finance companies.

ES3Interactive TV advertising and media technology. Since 2012.
Zone TVFree ad-supported streaming TV. Since 2016.
FioMobile health platform for community healthcare. Since 2023.
North Palm CapitalInvestment company. Since 2025.
ScrybPublic AI and analytics company. Since 2021.
TossdownE-commerce platform for the food industry. Since 2021.

Things I've built

Three tools of my own that run in production clusters and customer environments. I built all three with AI-assisted development.

CloudScope

Self-hosted AWS scanner for multiple accounts

Scans AWS accounts for security, cost, fault-tolerance and operational problems, and tracks billing through Cost Explorer: daily costs, anomalies, tag coverage and budget vs actual. It covers what AWS Trusted Advisor does without a Business Support plan, and runs as a few containers on EKS or ECS Fargate with PostgreSQL.

ik8sa

Intelligent Kubernetes agent, written in Go

Runs inside production clusters with its own backend and web UI. It uses AI to learn Kubernetes best practices, checks the cluster against them, and alerts on posture and configuration problems in real time. A standalone mode, so it can act on its own, is in progress.

Watchpost

Multi-tenant monitoring platform

My replacement for Zabbix and Datadog Synthetics. It has more than 11 check types, including declarative browser checks and HLS video-playback checks, plus customer-facing health dashboards and TOTP MFA. It runs on Fargate.

I use LLMs for architecture reasoning, code generation and incident diagnosis, and I treat prompting as an engineering skill like any other.

How an engagement works

I look first. I map what you run and what it costs, and find the biggest risks and the biggest waste.
I fix it live. Changes go in as code with zero-downtime rollouts, starting with whatever costs or hurts the most.
I stay on, if you want me to. Pipelines, monitoring and regular cost checks keep it working, on a retainer or when you call.

About me

I joined ES3 in 2012 to run its infrastructure and moved it from Rackspace to AWS in 2013. For the next few years that meant Windows servers, SQL Server reporting, and keeping ES3's consumer TV apps running: a games platform, Santa Tracker, and Facebook and Twitter for TV. In 2016 Zone TV started up in the same office as a separate company with its own product, a free streaming TV service. In 2018 I took over Linux management as well, moved into DevOps and AWS cloud work, and have run the infrastructure behind both products since. That includes moving two self-managed kops clusters to EKS in 2023, and ES3's whole workload into its own AWS account in 2025 without interrupting service. At ES3 alone that adds up to more than 40 major projects and over a thousand production deployments through the pipelines I run.

Day to day that means the whole stack: EKS clusters and the Bitbucket pipelines that deploy to them, the production launch of INGAGE, ES3's interactive TV advertising platform, and the data behind the reports. Right now I'm moving ES3's analytics database off Redshift onto PostgreSQL: 232 million rows and three and a half years of history, with dashboards switched over one at a time so there's always a way back.

A lot of my recent work is modernizing legacy services: rewriting old APIs on current frameworks and switching production over without downtime. I do much of that with Claude Code, which is also how I built CloudScope, Watchpost and several websites, including this one.

IT Delivered is how I offer the same work to other companies. Some keep me on a monthly retainer, some hire me for one project, and some call when something breaks. Either way you deal with me directly, and the person who plans the change is the one who makes it.

Before ES3 I spent seven years in network and systems engineering, then worked at Intel and IBM. In 2018 I also did digital TV engineering at Rogers, where I led the CraveTV launch across all markets. At Intel I drove vPro deployments for enterprise customers including BMW, Boeing, HSBC, Kraft Foods, Fuji Film and EDS, remote and on site across North America and EMEA, under formal SLAs and ISO processes.

Certifications: AWS Solutions Architect Associate, certified 2020
Microsoft SQL Server 2008, Implementation and Maintenance, 2012
Training: Linux Foundation Certified Systems Administrator (LFCS) course, Udemy, 2024
Kubernetes Essentials, A Cloud Guru, 2021
Introduction to Containers and Docker, Linux Academy, 2021
Database fundamentals, Seneca College, 2012-2013
Based in: Greater Toronto Area, and remote
Languages: English, Russian and Hebrew, working French